From violent at death.kiev.ua Wed Dec 1 13:41:50 2021 From: violent at death.kiev.ua (Igor Grabin) Date: Wed, 1 Dec 2021 13:41:50 +0200 Subject: [uanog] =?koi8-r?b?y9TPIMvByyDczMXL1NLJy9Ug097J1MHF1D8=?= Message-ID: <20211201114150.GR30278@death.kiev.ua> ???? ??????, ??????? ??????????, ????? ????????? ????? ???? ????? ????????. ? ???????, ??? ??? ?????? ??????? ????????. ?? ???? ????????? ??? ????????? ?????-?? ????? ?????, ???? ????????? baseline. ?? ????? ????? ?????? baseline. - ????? ????? ?????????? ????????? ?? ? ?? ???????? - ??? ??-??????????? ????????? ??????? ? 2??? mbr, -- Igor "CacoDem0n" Grabin From alex at euro.net.ua Wed Dec 1 15:57:18 2021 From: alex at euro.net.ua (Alexander V Soroka) Date: Wed, 1 Dec 2021 15:57:18 +0200 Subject: [uanog] =?koi8-r?b?y9TPIMvByyDczMXL1NLJy9Ug097J1MHF1D8=?= In-Reply-To: <20211201114150.GR30278@death.kiev.ua> References: <20211201114150.GR30278@death.kiev.ua> Message-ID: <668553619.20211201155718@euro.net.ua> ?????? ! ??? ???????? ???????? (???????????) ??????? ????? ???????????? ?? ???????????? ????????????? ??????????? ????? ??? ? ???? ??? ?????. ?? ? ????? ???-?? 10% ? ????? 20% :-) ?? "??????????". ???? ????? 2?? ?? ???? ????????: ??? ???????????? ? ?????? ????? ? ??? ??? "???????? ?????", ?.?. ???? ?? ????????? 100% ? ?????? ? ????????, ? ?????? ?????? ????? ??????? ? "?????? ????". ?.?. ?????? ?? ?????? ? ?????? ?? ? ????? ??? ?????????? ??????????. ??! ???? ? ???? ?????? "???? 1" ?? ??2 ?????????? ? ?????? ?????, ?.?. ????????? ???????? ?? 100% ?? ????. ? ???? ????? ????? ? ???? "??" ?? ??????? ???? ?? ?????? ?? ???? (????????) ??, ? ????????? ?????? 5%-10%. ??????: ? ?????? 10 ???????? ?? 1??? ????? ??1 ? 1??? ??2. ??????? ??? "????? ??????????" = 10*1???=10??? + 5-10%. ??????????????, ???? ???? ???? ? "?????????? ? "????1" ? "????2" ?? ???????? ?????????? ????1 = ???????? ?????????? ????2. ?.?. ??? 10??? ??????????? (????????) ? ????1 ?????????? ?? 10???+10% ? ? ????2 ?????????? ?? 10???+10%. Wednesday, December 1, 2021, 1:41:50 PM, Igor Grabin violent at death.kiev.ua you wrote: IG> ???? ??????, IG> ??????? ??????????, ????? ????????? ????? ???? ????? ????????. IG> ? ???????, ??? ??? ?????? ??????? ????????. ?? ???? ????????? ??? IG> ????????? ?????-?? ????? ?????, ???? ????????? baseline. IG> ?? ????? ????? ?????? baseline. IG> - ????? ????? ?????????? ????????? ?? ? ?? ???????? IG> - ??? ??-??????????? ????????? ??????? ? 2??? IG> mbr, -- Best regards, Alexander V Soroka http://www.svr.ua/ AS106-RIPE mailto:alex at euro.net.ua From vladimir.sharun at ukr.net Wed Dec 1 16:48:51 2021 From: vladimir.sharun at ukr.net (Vladimir Sharun) Date: Wed, 01 Dec 2021 16:48:51 +0200 Subject: [uanog] =?utf-8?b?0LrRgtC+INC60LDQuiDRjdC70LXQutGC0YDQuNC60YMg?= =?utf-8?b?0YHRh9C40YLQsNC10YI/?= In-Reply-To: <20211201114150.GR30278@death.kiev.ua> References: <20211201114150.GR30278@death.kiev.ua> Message-ID: <1638367837.223895000.4vraprqb@frv53.fwdcdn.com> ??????, ? ???? ????? ? ??????? ?????, ??????????? ? ????? ?? ? ??? ? ????? ????? ??????? ? ?????? ???????? ??????? ???? ?-? ?-? ?????????? ????????, ?????? ??? ?? ?? ??????????? ????????????? ?????? ???????, ? ???????????? ??????????? ?? ?????? - ??? ???? ???????????. ???? ?? ???????? ??????????? ???? - 85??, ???? ????? ??????? ???? ?? ???????? (? ??? ??? ????? ???????? ?????/????? ???????? ???? ?? ?????? ????? ???? ??????), ?????? "?? ????, ????? ?????", ? ?? ???? ????? ??????? "? ??? ?????????, ??? ???????, ??? 85??, ???????? ????". ???????? ?? ????????, ???? ?????. ??????? ??? 85?? - ??? 50% ???????????? ???????? ? 50% ????????? ?? ?????? ? ??????. ??, ??????? ?????????? ??? ????????????. ??????????? ?? ??? ?????? ?? ???????? ??????????? ? ?????? ???? ???-?? ???????? ???????????? ???????? - ??? ????? ? BMC. ??? ?? ??????????? ?????????????, ?? ???? ?? ?? ??? ?? ????????????? ????? ???? ??????? ?? ????? Xeon Scalable/EPYC, ?? ???-?? ??????/???-?? ?????? ? 85?? ?? ???? ??? ????????? ???????? ?? ???????? (3.5-3.75??? ?? ??????). ?? ?????? ?? 5???, ?? ?????? ??????? ??????? ??????????? "? ????? ??? ??????? ? ????? ????????? ?? ?????" ? ????????????????? ????? ????? ??????. ???? ? ???? ????? ????? ????????, ?? ??????????? ????? ???? ?? 10-15% ????.? ?????????? ?????? ??????? ????? ?????????? ?? ???? ???????? :-) ? ??, ?????? ??? ???? ?????? - ??????? ?? ????? ??????. ?? ???? ?? ???????. 1 ?????? 2021, 13:42:01, ??? "Igor Grabin" : ???? ??????, ??????? ??????????, ????? ????????? ????? ???? ????? ????????. ? ???????, ??? ??? ?????? ??????? ????????. ?? ???? ????????? ??? ????????? ?????-?? ????? ?????, ???? ????????? baseline. ?? ????? ????? ?????? baseline. - ????? ????? ?????????? ????????? ?? ? ?? ???????? - ??? ??-??????????? ????????? ??????? ? 2??? mbr, -- Igor "CacoDem0n" Grabin _______________________________________________ uanog mailing list uanog at uanog.kiev.ua https://mailman.uanog.kiev.ua/mailman/listinfo/uanog -------------- next part -------------- An HTML attachment was scrubbed... URL: From basil at vpm.net.ua Wed Dec 1 17:23:35 2021 From: basil at vpm.net.ua (VASYL MELNYK) Date: Wed, 1 Dec 2021 17:23:35 +0200 Subject: [uanog] =?utf-8?b?0LrRgtC+INC60LDQuiDRjdC70LXQutGC0YDQuNC60YMg?= =?utf-8?b?0YHRh9C40YLQsNC10YI/?= In-Reply-To: <1638367837.223895000.4vraprqb@frv53.fwdcdn.com> References: <20211201114150.GR30278@death.kiev.ua> <1638367837.223895000.4vraprqb@frv53.fwdcdn.com> Message-ID: > > ??????? ??? 85?? - ??? 50% ???????????? ???????? ? 50% ????????? ?? ?????? > ? ??????. > ?????????, ???????? ??? ? ????? ??? ????? ????? ?????? ?????? ?? ????: ?????????????? ??????? ?? ???????????? ??? ????????????? ?? ????????? ? ?? ??? ??? ??????????? ?????? -------------- next part -------------- An HTML attachment was scrubbed... URL: From vladimir.sharun at ukr.net Wed Dec 1 18:56:32 2021 From: vladimir.sharun at ukr.net (Vladimir Sharun) Date: Wed, 01 Dec 2021 18:56:32 +0200 Subject: [uanog] =?utf-8?b?0LrRgtC+INC60LDQuiDRjdC70LXQutGC0YDQuNC60YMg?= =?utf-8?b?0YHRh9C40YLQsNC10YI/?= In-Reply-To: References: <20211201114150.GR30278@death.kiev.ua> <1638367837.223895000.4vraprqb@frv53.fwdcdn.com> Message-ID: <1638377421.795529000.hxnqnlee@frv53.fwdcdn.com> ??????, ??? ????????? ????? - ???? ???, ?? ? ?? ??????, ?????? ???: ??? ?????????? ???? ????-????? Feed A & Feed B ????? ???? ????? ???? ????? ?????? ????? ??????? ?? ?????? UPS, ? ?? ?? ??????????? ??????? .... ? ????? ??? ?????? ?? ?????????, ??????? ??????? ????? ??????? ????????? (??????????) ??? ????? ????????????-?????? "?????????" Tier1 ?????? ???? ??? ? ??????? 1 ?????? 2021, 17:21:53, ??? "VASYL MELNYK" : ??????? ??? 85?? - ??? 50% ???????????? ???????? ? 50% ????????? ?? ?????? ? ??????. ?????????, ???????? ??? ? ????? ??? ????? ????? ?????? ?????? ?? ????: ?????????????? ??????? ?? ???????????? ??? ????????????? ?? ????????? ? ?? ??? ??? ??????????? ?????? -------------- next part -------------- An HTML attachment was scrubbed... URL: From basil at vpm.net.ua Wed Dec 1 19:01:18 2021 From: basil at vpm.net.ua (VASYL MELNYK) Date: Wed, 1 Dec 2021 19:01:18 +0200 Subject: [uanog] =?utf-8?b?0LrRgtC+INC60LDQuiDRjdC70LXQutGC0YDQuNC60YMg?= =?utf-8?b?0YHRh9C40YLQsNC10YI/?= In-Reply-To: <1638377421.795529000.hxnqnlee@frv53.fwdcdn.com> References: <20211201114150.GR30278@death.kiev.ua> <1638367837.223895000.4vraprqb@frv53.fwdcdn.com> <1638377421.795529000.hxnqnlee@frv53.fwdcdn.com> Message-ID: ? ????? ?? ?????? ???? ?????????? ?? ????????? - ???????? ??? ? ?????. ????? ??? ???????? ?? ?????? ?????, ?? ???? ?????? ???????, ? ?????? ??????? ????? ???????? ?? ?????? ???? - ??? ???. ??, 1 ???. 2021 ?. ? 18:56, Vladimir Sharun : > ??????, > > ??? ????????? ????? - ???? ???, ?? ? ?? ??????, ?????? ???: > > - ??? ?????????? ???? ????-????? > - Feed A & Feed B ????? ???? ????? ???? ????? ?????? ????? ??????? ?? > ?????? UPS, ? ?? ?? ??????????? ??????? > - .... ? ????? ??? ?????? ?? ?????????, ??????? ??????? ????? > ??????? ????????? (??????????) ??? ????? ????????????-?????? > - "?????????" Tier1 ?????? ???? ??? ? ??????? > > > *1 ?????? 2021, 17:21:53, ??? "VASYL MELNYK" >: * > > ??????? ??? 85?? - ??? 50% ???????????? ???????? ? 50% ????????? ?? ?????? > ? ??????. > > > ?????????, ???????? ??? ? ????? > > ??? ????? ????? ?????? ?????? ?? ????: ?????????????? ??????? ?? > ???????????? ??? ????????????? ?? ????????? ? ?? ??? ??? ??????????? ?????? > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From oleh.hrynchuk at gmail.com Thu Dec 2 09:51:19 2021 From: oleh.hrynchuk at gmail.com (Oleh Hrynchuk) Date: Thu, 2 Dec 2021 09:51:19 +0200 Subject: [uanog] =?utf-8?b?0KXRgtC+INGJ0L7RgdGMINGF0L7Rh9C1INC/0L7QvNGW?= =?utf-8?b?0L3Rj9GC0Lgg0LIg0LbQuNGC0YLRli/Qv9C+0YfQsNGC0Lgg0YHQv9C+?= =?utf-8?b?0YfQsNGC0LrRgyA/IDop?= Message-ID: ?????? ???? (????? ??????? ? ???????, ???????...) ???????? ???????? ? ??????. ??? ?? ???? ?????? ?? ?? CTO, ?? ?? ???? ???????... ????? ??????-????? ??? ??????. https://www.pracuj.pl/praca/specjalista-ds-it-stargard,oferta,1001336790 -- Regards, /oleh hrynchuk -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: ?? ?????!.jpg Type: image/jpeg Size: 39603 bytes Desc: not available URL: From alexander.moskalenko at gmail.com Mon Dec 6 16:00:57 2021 From: alexander.moskalenko at gmail.com (Oleksandr Moskalenko) Date: Mon, 6 Dec 2021 15:00:57 +0100 Subject: [uanog] =?utf-8?b?0KXRgtC+INGJ0L7RgdGMINGF0L7Rh9C1INC/0L7QvNGW?= =?utf-8?b?0L3Rj9GC0Lgg0LIg0LbQuNGC0YLRli/Qv9C+0YfQsNGC0Lgg0YHQv9C+?= =?utf-8?b?0YfQsNGC0LrRgyA/IDop?= In-Reply-To: References: Message-ID: ????? ?? ????? ? ??? :) https://amazon.jobs/en/jobs/1810788/senior-technical-account-manager-enterprise-engineer ? ??? ????? ?????? ??????? On Thu, Dec 2, 2021 at 8:51 AM Oleh Hrynchuk wrote: > ?????? ???? (????? ??????? ? ???????, ???????...) ???????? ???????? ? > ??????. > ??? ?? ???? ?????? ?? ?? CTO, ?? ?? ???? ???????... ????? ??????-????? ??? > ??????. > > https://www.pracuj.pl/praca/specjalista-ds-it-stargard,oferta,1001336790 > > -- > Regards, > /oleh hrynchuk > _______________________________________________ > uanog mailing list > uanog at uanog.kiev.ua > https://mailman.uanog.kiev.ua/mailman/listinfo/uanog -------------- next part -------------- An HTML attachment was scrubbed... URL: From oleh.hrynchuk at gmail.com Mon Dec 6 18:50:31 2021 From: oleh.hrynchuk at gmail.com (Oleh Hrynchuk) Date: Mon, 6 Dec 2021 18:50:31 +0200 Subject: [uanog] =?utf-8?b?0KXRgtC+INGJ0L7RgdGMINGF0L7Rh9C1INC/0L7QvNGW?= =?utf-8?b?0L3Rj9GC0Lgg0LIg0LbQuNGC0YLRli/Qv9C+0YfQsNGC0Lgg0YHQv9C+?= =?utf-8?b?0YfQsNGC0LrRgyA/IDop?= In-Reply-To: References: Message-ID: ??, ?? ????? ?? ????????????, ?? ? ??????? ???????????. Really! ??? ?? ??????? ??? AWS-?????? ?? ?????? ???? ??, 6 ????. 2021 ? 16:01 Oleksandr Moskalenko < alexander.moskalenko at gmail.com> ????: > ????? ?? ????? ? ??? :) > > > https://amazon.jobs/en/jobs/1810788/senior-technical-account-manager-enterprise-engineer > ? ??? ????? ?????? ??????? > > > On Thu, Dec 2, 2021 at 8:51 AM Oleh Hrynchuk > wrote: > >> ?????? ???? (????? ??????? ? ???????, ???????...) ???????? ???????? ? >> ??????. >> ??? ?? ???? ?????? ?? ?? CTO, ?? ?? ???? ???????... ????? ??????-????? >> ??? ??????. >> >> https://www.pracuj.pl/praca/specjalista-ds-it-stargard,oferta,1001336790 >> >> -- >> Regards, >> /oleh hrynchuk >> _______________________________________________ >> uanog mailing list >> uanog at uanog.kiev.ua >> https://mailman.uanog.kiev.ua/mailman/listinfo/uanog > > -- Regards, /oleh hrynchuk -------------- next part -------------- An HTML attachment was scrubbed... URL: From netch at netch.kiev.ua Mon Dec 6 19:12:00 2021 From: netch at netch.kiev.ua (Valentin Nechayev) Date: Mon, 6 Dec 2021 19:12:00 +0200 Subject: [uanog] =?utf-8?b?0KXRgtC+INGJ0L7RgdGMINGF0L7Rh9C1INC/0L7QvNGW?= =?utf-8?b?0L3Rj9GC0Lgg0LIg0LbQuNGC0YLRli/Qv9C+0YfQsNGC0Lgg0YHQv9C+0Yc=?= =?utf-8?b?0LDRgtC60YMgPyA6KQ==?= In-Reply-To: References: Message-ID: <20211206171200.GQ145444@netch.kiev.ua> hi, Mon, Dec 06, 2021 at 18:50:31, oleh.hrynchuk wrote about "Re: [uanog] ??? ???? ???? ???????? ? ?????/?????? ???????? ? :)": > ??, ?? ????? ?? ????????????, ?? ? ??????? ???????????. > Really! > ??? ?? ??????? ??? AWS-?????? ??? ?????? ?? DOU - "???????? ? FAANG, ? ???????? ?????????? ? GAFNA" -netch- From doka at funlab.cc Tue Dec 28 11:01:50 2021 From: doka at funlab.cc (Volodymyr Litovka) Date: Tue, 28 Dec 2021 11:01:50 +0200 Subject: [uanog] switch's port buffers In-Reply-To: References: Message-ID: ??????, ??? ????????? ?????????, ???? ?????????? ?????? ??? ????????? ??????????? ? ????????? ???????? ?? ??????????? ???????? ? ???????????? ???????????? ? 2019 ???? - "*It hosted 98 attendees from 12 economies with 26 from academia and 72 from industry.*" -- https://blog.apnic.net/2019/12/12/sizing-the-buffer/ ??? ?????? ?? ????????? ?????? - ????????? ????? ???????? ???????? ???: http://buffer-workshop.stanford.edu/program/ > [ ... ] buffers also add additional lag to a packet?s transit through > the network. If you want to implement a low jitter service, then deep > buffers are decidedly unfriendly! The result is the rather enigmatic > observation that network buffers have to be as big as they need to be, > but no bigger! ??? ?????? ??? :) ??????? ?????? ?? ??????: === 1. ??????????? "????? 40MB - ??????" ?? ??????????? ???????? ?????????????? ?????????????? === > A study by a Stanford TCP research group in 2004 > used the central limit > theorem to point to a radically smaller model of buffer size. Link > efficiency can be maintained for N desynchronized flows with a buffer > that is dimensioned to the size of: > > *Size*?= (*/BW/*?? */RTT/*) / ?*/N/* > > This is a radical result for high-speed extended latency links in a > busy network. The consequences on router design are enormous: > > ?For example, a 1 Tb/s ISP router carrying one TCP flow with an > RTTmin of 100ms would require 12.5 GB of buffer and off-chip > buffering. *If it carries 100,000 flows, then the buffer can be > safely reduced to less than 40MB, reducing the buffering and > worst-case latency by 99.7%*. With small buffers, the buffer would > comfortably fit on a single chip switch ASIC.? > > Nick McKeown et. al. Sizing Router Buffers (Redux) > > ????? ???????, ?????????? ?? ???????? ??????????? ???????????? ?? ?????? https://people.ucsc.edu/~warner/buffer.html ? ???????? ????????? 48x10/25 + 6-8/100 ?????????????? ???????? ?????? 32-40MB - ????? ??????????. ? ???? ??????, ??? ???????????? ?????? - ??? 1/10G ????????? (???? ?? 25), ?? ????? ????????????, ??? ???????? is absolutely safe ? ??????? ?????? ??????? ? ????? ????????????. === 2. ????????????? inline notification ???????? ???????? ???????? ??????? === > The advantage of ECN is that the sender is not placed in the position > of being informed of a congestion condition well after the condition > has occurred. Explicit notification allows the sender to be informed > of a condition as it is forming so that it can take action while there > is still a coherent ack pacing signal coming back from the receiver > (that is before packet loss occurs). > > However, ECN is only a single bit marking. Is this enough? [ ... ] The > conclusion from one presentation is that the single-bit marking, while > coarse and non-specific is probably sufficient to moderate > self-clocking TCP flows such that they do not place pressure on > network buffers, leaving the buffers to deal with short term bursts > from unconstrained sources. > [ ... ] > > And if we want to reduce buffer size and maintain efficient and fair > performance how can we achieve it? One view is that sender pacing can > remove much of the pressure on buffers, and self-clocking flows can > stabilise without emitting transient bursts that need to be absorbed > by buffers. Another view, and one that does not necessarily contradict > the first, is that the self-clocking algorithm can operate with higher > precision if there was some form of feedback from the network on the > state of the network path. This can be as simple as a single bit (ECN) > or a complete trace of path element queue state (HPCC). > ??? ?? ????? ????? ?????? ????????? ECN === 3. flow-aware traffic management ?? ???????? ??????????? ??????? ?? ????? === > If the network was in a position to be able to classify all currently > active flows into either elephants or mice, then the network could be > able to use different queuing regimes for each traffic class. This > sorting adds to the cost and complexity of packet switches, and if > scaling pressures are a factor in switch design then it?s not clear > that the additional cost of switch complexity would be offset by a far > superior efficiency outcome in the switching function. === 4. ? ???? ????? ??????? ???????? ???????? ??? ?????????? (???? 100ms) ?????? === > Further analysis reveals an estimate of packet drop rates if the > network?s buffers were reduced in size, and for this particular case, > the analysis revealed that an 18msec buffer would be able to sustain a > packet drop rate of less than 0.005%. ???????. On 15.10.2021 14:48, Volodymyr Litovka wrote: > > ??????, > > ? ??? ??? ???? ????????, ? ? ?? ???? ?? ?????????? :) ? ???????, ?? > ??????? ???? ??????? ?????? ?????, ??? ??? ?? ???? - > > ?? ????? ????????, ??????? ?????? (?? ???? ??? shared) ????????? ??? > ????????? ??? ???????? ?? ??????????? ????? ????? ??????? ??? ????? > traffic bursts? ?? ????? ????????? ?? ?????????? ?????????? ????? > ?????????? ?? ??????? ?? ?????? "? ???? ??? ???????", "? ??????? > ??????? ????? ? ????????? ?? ???????????" ?? ???? ????, ? ?? > ?????????-??????????? - ???????? ?????? ? ????? ?????? ? ???????? > ?????? ????????? ?????????? ??, ?????????, ?? ??????????? ?????????? > 24x1G + 2x10G ? ???????????? ????????????????????? ??????????? > > ????? ??????? ??? ???????? ????????? :) > > ?????. > > -- > Volodymyr Litovka > "Vision without Execution is Hallucination." -- Thomas Edison -- Volodymyr Litovka "Vision without Execution is Hallucination." -- Thomas Edison -------------- next part -------------- An HTML attachment was scrubbed... URL: From maxtul at netassist.kiev.ua Tue Dec 28 11:26:31 2021 From: maxtul at netassist.kiev.ua (Max Tulyev) Date: Tue, 28 Dec 2021 11:26:31 +0200 Subject: [uanog] switch's port buffers In-Reply-To: References: Message-ID: <7c8d121e-64a9-9be0-5ba3-c9a2c3632f75@netassist.kiev.ua> ??????! ????? ?????????????, ???????! ?????? ???? ??????: ?????? ?? ???????? ethernet flow control ?????????? ????? ?????????? ????? ????? ????? ???????? ? ?????? ???????. 28.12.21 11:01, Volodymyr Litovka ????: > ??????, > > ??? ????????? ?????????, ???? ?????????? ?????? ??? ????????? > ??????????? ? ????????? ???????? ?? ??????????? ???????? ? ???????????? > ???????????? ? 2019 ???? - "*It hosted 98 attendees from 12 economies > with 26 from academia and 72 from industry.*" -- > https://blog.apnic.net/2019/12/12/sizing-the-buffer/ > > ??? ?????? ?? ????????? ?????? - ????????? ????? ???????? ???????? ???: > http://buffer-workshop.stanford.edu/program/ > >> [ ... ] buffers also add additional lag to a packet?s transit through >> the network. If you want to implement a low jitter service, then deep >> buffers are decidedly unfriendly! The result is the rather enigmatic >> observation that network buffers have to be as big as they need to be, >> but no bigger! > > ??? ?????? ??? :) ??????? ?????? ?? ??????: > > === 1. ??????????? "????? 40MB - ??????" ?? ??????????? ???????? > ?????????????? ?????????????? === > >> A study by a Stanford TCP research group in 2004 >> used the central limit >> theorem to point to a radically smaller model of buffer size. Link >> efficiency can be maintained for N desynchronized flows with a buffer >> that is dimensioned to the size of: >> >> *Size*?= (*/BW/*?? */RTT/*) / ?*/N/* >> >> This is a radical result for high-speed extended latency links in a >> busy network. The consequences on router design are enormous: >> >> ?For example, a 1 Tb/s ISP router carrying one TCP flow with an >> RTTmin of 100ms would require 12.5 GB of buffer and off-chip >> buffering. *If it carries 100,000 flows, then the buffer can be >> safely reduced to less than 40MB, reducing the buffering and >> worst-case latency by 99.7%*. With small buffers, the buffer would >> comfortably fit on a single chip switch ASIC.? >> >> Nick McKeown et. al. Sizing Router Buffers (Redux) >> >> > ????? ???????, ?????????? ?? ???????? ??????????? ???????????? ?? ?????? > https://people.ucsc.edu/~warner/buffer.html ? ???????? ????????? > 48x10/25 + 6-8/100 ?????????????? ???????? ?????? 32-40MB - ????? > ??????????. ? ???? ??????, ??? ???????????? ?????? - ??? 1/10G ????????? > (???? ?? 25), ?? ????? ????????????, ??? ???????? is absolutely safe ? > ??????? ?????? ??????? ? ????? ????????????. > > === 2. ????????????? inline notification ???????? ???????? ???????? > ??????? === > >> The advantage of ECN is that the sender is not placed in the position >> of being informed of a congestion condition well after the condition >> has occurred. Explicit notification allows the sender to be informed >> of a condition as it is forming so that it can take action while there >> is still a coherent ack pacing signal coming back from the receiver >> (that is before packet loss occurs). >> >> However, ECN is only a single bit marking. Is this enough? [ ... ] The >> conclusion from one presentation is that the single-bit marking, while >> coarse and non-specific is probably sufficient to moderate >> self-clocking TCP flows such that they do not place pressure on >> network buffers, leaving the buffers to deal with short term bursts >> from unconstrained sources. >> > [ ... ] >> >> And if we want to reduce buffer size and maintain efficient and fair >> performance how can we achieve it? One view is that sender pacing can >> remove much of the pressure on buffers, and self-clocking flows can >> stabilise without emitting transient bursts that need to be absorbed >> by buffers. Another view, and one that does not necessarily contradict >> the first, is that the self-clocking algorithm can operate with higher >> precision if there was some form of feedback from the network on the >> state of the network path. This can be as simple as a single bit (ECN) >> or a complete trace of path element queue state (HPCC). >> > > ??? ?? ????? ????? ?????? ????????? ECN > > === 3. flow-aware traffic management ?? ???????? ??????????? ??????? ?? > ????? === > >> If the network was in a position to be able to classify all currently >> active flows into either elephants or mice, then the network could be >> able to use different queuing regimes for each traffic class. This >> sorting adds to the cost and complexity of packet switches, and if >> scaling pressures are a factor in switch design then it?s not clear >> that the additional cost of switch complexity would be offset by a far >> superior efficiency outcome in the switching function. > > === 4. ? ???? ????? ??????? ???????? ???????? ??? ?????????? (???? > 100ms) ?????? === > >> Further analysis reveals an estimate of packet drop rates if the >> network?s buffers were reduced in size, and for this particular case, >> the analysis revealed that an 18msec buffer would be able to sustain a >> packet drop rate of less than 0.005%. > > ???????. > > > On 15.10.2021 14:48, Volodymyr Litovka wrote: >> >> ??????, >> >> ? ??? ??? ???? ????????, ? ? ?? ???? ?? ?????????? :) ? ???????, ?? >> ??????? ???? ??????? ?????? ?????, ??? ??? ?? ???? - >> >> ?? ????? ????????, ??????? ?????? (?? ???? ??? shared) ????????? ??? >> ????????? ??? ???????? ?? ??????????? ????? ????? ??????? ??? ????? >> traffic bursts? ?? ????? ????????? ?? ?????????? ?????????? ????? >> ?????????? ?? ??????? ?? ?????? "? ???? ??? ???????", "? ??????? >> ??????? ????? ? ????????? ?? ???????????" ?? ???? ????, ? ?? >> ?????????-??????????? - ???????? ?????? ? ????? ?????? ? ???????? >> ?????? ????????? ?????????? ??, ?????????, ?? ??????????? ?????????? >> 24x1G + 2x10G ? ???????????? ????????????????????? ??????????? >> >> ????? ??????? ??? ???????? ????????? :) >> >> ?????. >> >> -- >> Volodymyr Litovka >> "Vision without Execution is Hallucination." -- Thomas Edison > > -- > Volodymyr Litovka > "Vision without Execution is Hallucination." -- Thomas Edison > > > _______________________________________________ > uanog mailing list > uanog at uanog.kiev.ua > https://mailman.uanog.kiev.ua/mailman/listinfo/uanog From doka at funlab.cc Tue Dec 28 11:36:39 2021 From: doka at funlab.cc (Volodymyr Litovka) Date: Tue, 28 Dec 2021 11:36:39 +0200 Subject: [uanog] switch's port buffers In-Reply-To: <7c8d121e-64a9-9be0-5ba3-c9a2c3632f75@netassist.kiev.ua> References: <7c8d121e-64a9-9be0-5ba3-c9a2c3632f75@netassist.kiev.ua> Message-ID: <579150fd-e9c1-90e5-cfce-89d842dbc761@funlab.cc> ??????, ???? ????????????, ??? ??? ??-?? ?????????? ???????? ????? ?? ??????? IP. ??????????? ?????????? ???????? ??????, ?????? ?? ???? ? ????????? underlay ????. ?????????? ? ?????? ECN - ??? ??????? IP ? ?????????? ?????????? ???????????? ?????? IP-?????????, ??????????? ????? ??????? ???????? ????? ?? ??????? ????. ???????? ??? ???????? ????? ???????? ? ???????????????? ? ??????????????? ????????????. On 28.12.2021 11:26, Max Tulyev wrote: > ??????! > > ????? ?????????????, ???????! > > ?????? ???? ??????: ?????? ?? ???????? ethernet flow control > ?????????? ????? ?????????? ????? ????? ????? ???????? ? ?????? ???????. > > 28.12.21 11:01, Volodymyr Litovka ????: >> ??????, >> >> ??? ????????? ?????????, ???? ?????????? ?????? ??? ????????? >> ??????????? ? ????????? ???????? ?? ??????????? ???????? ? >> ???????????? ???????????? ? 2019 ???? - "*It hosted 98 attendees from >> 12 economies with 26 from academia and 72 from industry.*" -- >> https://blog.apnic.net/2019/12/12/sizing-the-buffer/ >> >> ??? ?????? ?? ????????? ?????? - ????????? ????? ???????? ???????? >> ???: http://buffer-workshop.stanford.edu/program/ >> >>> [ ... ] buffers also add additional lag to a packet?s transit >>> through the network. If you want to implement a low jitter service, >>> then deep buffers are decidedly unfriendly! The result is the rather >>> enigmatic observation that network buffers have to be as big as they >>> need to be, but no bigger! >> >> ??? ?????? ??? :) ??????? ?????? ?? ??????: >> >> === 1. ??????????? "????? 40MB - ??????" ?? ??????????? ???????? >> ?????????????? ?????????????? === >> >>> A study by a Stanford TCP research group in 2004 >>> used the central limit >>> theorem to point to a radically smaller model of buffer size. Link >>> efficiency can be maintained for N desynchronized flows with a >>> buffer that is dimensioned to the size of: >>> >>> *Size*?= (*/BW/*?? */RTT/*) / ?*/N/* >>> >>> This is a radical result for high-speed extended latency links in a >>> busy network. The consequences on router design are enormous: >>> >>> ??? ?For example, a 1 Tb/s ISP router carrying one TCP flow with an >>> ??? RTTmin of 100ms would require 12.5 GB of buffer and off-chip >>> ??? buffering. *If it carries 100,000 flows, then the buffer can be >>> ??? safely reduced to less than 40MB, reducing the buffering and >>> ??? worst-case latency by 99.7%*. With small buffers, the buffer would >>> ??? comfortably fit on a single chip switch ASIC.? >>> >>> ??? Nick McKeown et. al. Sizing Router Buffers (Redux) >>> >>> >> ????? ???????, ?????????? ?? ???????? ??????????? ???????????? ?? >> ?????? https://people.ucsc.edu/~warner/buffer.html ? ???????? >> ????????? 48x10/25 + 6-8/100 ?????????????? ???????? ?????? 32-40MB - >> ????? ??????????. ? ???? ??????, ??? ???????????? ?????? - ??? 1/10G >> ????????? (???? ?? 25), ?? ????? ????????????, ??? ???????? is >> absolutely safe ? ??????? ?????? ??????? ? ????? ????????????. >> >> === 2. ????????????? inline notification ???????? ???????? ???????? >> ??????? === >> >>> The advantage of ECN is that the sender is not placed in the >>> position of being informed of a congestion condition well after the >>> condition has occurred. Explicit notification allows the sender to >>> be informed of a condition as it is forming so that it can take >>> action while there is still a coherent ack pacing signal coming back >>> from the receiver (that is before packet loss occurs). >>> >>> However, ECN is only a single bit marking. Is this enough? [ ... ] >>> The conclusion from one presentation is that the single-bit marking, >>> while coarse and non-specific is probably sufficient to moderate >>> self-clocking TCP flows such that they do not place pressure on >>> network buffers, leaving the buffers to deal with short term bursts >>> from unconstrained sources. >>> >> [ ... ] >>> >>> And if we want to reduce buffer size and maintain efficient and fair >>> performance how can we achieve it? One view is that sender pacing >>> can remove much of the pressure on buffers, and self-clocking flows >>> can stabilise without emitting transient bursts that need to be >>> absorbed by buffers. Another view, and one that does not necessarily >>> contradict the first, is that the self-clocking algorithm can >>> operate with higher precision if there was some form of feedback >>> from the network on the state of the network path. This can be as >>> simple as a single bit (ECN) or a complete trace of path element >>> queue state (HPCC). >>> >> >> ??? ?? ????? ????? ?????? ????????? ECN >> >> === 3. flow-aware traffic management ?? ???????? ??????????? ??????? >> ?? ????? === >> >>> If the network was in a position to be able to classify all >>> currently active flows into either elephants or mice, then the >>> network could be able to use different queuing regimes for each >>> traffic class. This sorting adds to the cost and complexity of >>> packet switches, and if scaling pressures are a factor in switch >>> design then it?s not clear that the additional cost of switch >>> complexity would be offset by a far superior efficiency outcome in >>> the switching function. >> >> === 4. ? ???? ????? ??????? ???????? ???????? ??? ?????????? (???? >> 100ms) ?????? === >> >>> Further analysis reveals an estimate of packet drop rates if the >>> network?s buffers were reduced in size, and for this particular >>> case, the analysis revealed that an 18msec buffer would be able to >>> sustain a packet drop rate of less than 0.005%. >> >> ???????. >> >> >> On 15.10.2021 14:48, Volodymyr Litovka wrote: >>> >>> ??????, >>> >>> ? ??? ??? ???? ????????, ? ? ?? ???? ?? ?????????? :) ? ???????, ?? >>> ??????? ???? ??????? ?????? ?????, ??? ??? ?? ???? - >>> >>> ?? ????? ????????, ??????? ?????? (?? ???? ??? shared) ????????? ??? >>> ????????? ??? ???????? ?? ??????????? ????? ????? ??????? ??? ????? >>> traffic bursts? ?? ????? ????????? ?? ?????????? ?????????? ????? >>> ?????????? ?? ??????? ?? ?????? "? ???? ??? ???????", "? ??????? >>> ??????? ????? ? ????????? ?? ???????????" ?? ???? ????, ? ?? >>> ?????????-??????????? - ???????? ?????? ? ????? ?????? ? ???????? >>> ?????? ????????? ?????????? ??, ?????????, ?? ??????????? ?????????? >>> 24x1G + 2x10G ? ???????????? ????????????????????? ??????????? >>> >>> ????? ??????? ??? ???????? ????????? :) >>> >>> ?????. >>> >>> -- >>> Volodymyr Litovka >>> ?? "Vision without Execution is Hallucination." -- Thomas Edison >> >> -- >> Volodymyr Litovka >> ?? "Vision without Execution is Hallucination." -- Thomas Edison >> >> >> _______________________________________________ >> uanog mailing list >> uanog at uanog.kiev.ua >> https://mailman.uanog.kiev.ua/mailman/listinfo/uanog > _______________________________________________ > uanog mailing list > uanog at uanog.kiev.ua > https://mailman.uanog.kiev.ua/mailman/listinfo/uanog -- Volodymyr Litovka "Vision without Execution is Hallucination." -- Thomas Edison -------------- next part -------------- An HTML attachment was scrubbed... URL: From maxtul at netassist.kiev.ua Tue Dec 28 12:00:07 2021 From: maxtul at netassist.kiev.ua (Max Tulyev) Date: Tue, 28 Dec 2021 12:00:07 +0200 Subject: [uanog] switch's port buffers In-Reply-To: <579150fd-e9c1-90e5-cfce-89d842dbc761@funlab.cc> References: <7c8d121e-64a9-9be0-5ba3-c9a2c3632f75@netassist.kiev.ua> <579150fd-e9c1-90e5-cfce-89d842dbc761@funlab.cc> Message-ID: <5778237a-43b1-1c37-799c-1ad0079973e6@netassist.kiev.ua> ??? ????????, ?????? pause ? ????????????? ??? ????, ????? ?????? ????? ? ????????? ?????????? ????! ?? ??? ?? ????????. ??????, ? ?????, ??? ??? ???? ???-?? ??? ?????????, ?? ???? ? ??? ????? ????? ??? ????? ? ???? TCP - ??? ?? ???, ???????? ??????????. 28.12.21 11:36, Volodymyr Litovka ????: > ??????, > > ???? ????????????, ??? ??? ??-?? ?????????? ???????? ????? ?? ??????? > IP. ??????????? ?????????? ???????? ??????, ?????? ?? ???? ? ????????? > underlay ????. ?????????? ? ?????? ECN - ??? ??????? IP ? ?????????? > ?????????? ???????????? ?????? IP-?????????, ??????????? ????? ??????? > ???????? ????? ?? ??????? ????. ???????? ??? ???????? ????? ???????? ? > ???????????????? ? ??????????????? ????????????. > > On 28.12.2021 11:26, Max Tulyev wrote: >> ??????! >> >> ????? ?????????????, ???????! >> >> ?????? ???? ??????: ?????? ?? ???????? ethernet flow control >> ?????????? ????? ?????????? ????? ????? ????? ???????? ? ?????? ???????. >> >> 28.12.21 11:01, Volodymyr Litovka ????: >>> ??????, >>> >>> ??? ????????? ?????????, ???? ?????????? ?????? ??? ????????? >>> ??????????? ? ????????? ???????? ?? ??????????? ???????? ? >>> ???????????? ???????????? ? 2019 ???? - "*It hosted 98 attendees from >>> 12 economies with 26 from academia and 72 from industry.*" -- >>> https://blog.apnic.net/2019/12/12/sizing-the-buffer/ >>> >>> ??? ?????? ?? ????????? ?????? - ????????? ????? ???????? ???????? >>> ???: http://buffer-workshop.stanford.edu/program/ >>> >>>> [ ... ] buffers also add additional lag to a packet?s transit >>>> through the network. If you want to implement a low jitter service, >>>> then deep buffers are decidedly unfriendly! The result is the rather >>>> enigmatic observation that network buffers have to be as big as they >>>> need to be, but no bigger! >>> >>> ??? ?????? ??? :) ??????? ?????? ?? ??????: >>> >>> === 1. ??????????? "????? 40MB - ??????" ?? ??????????? ???????? >>> ?????????????? ?????????????? === >>> >>>> A study by a Stanford TCP research group in 2004 >>>> used the central limit >>>> theorem to point to a radically smaller model of buffer size. Link >>>> efficiency can be maintained for N desynchronized flows with a >>>> buffer that is dimensioned to the size of: >>>> >>>> *Size*?= (*/BW/*?? */RTT/*) / ?*/N/* >>>> >>>> This is a radical result for high-speed extended latency links in a >>>> busy network. The consequences on router design are enormous: >>>> >>>> ??? ?For example, a 1 Tb/s ISP router carrying one TCP flow with an >>>> ??? RTTmin of 100ms would require 12.5 GB of buffer and off-chip >>>> ??? buffering. *If it carries 100,000 flows, then the buffer can be >>>> ??? safely reduced to less than 40MB, reducing the buffering and >>>> ??? worst-case latency by 99.7%*. With small buffers, the buffer would >>>> ??? comfortably fit on a single chip switch ASIC.? >>>> >>>> ??? Nick McKeown et. al. Sizing Router Buffers (Redux) >>>> >>>> >>> ????? ???????, ?????????? ?? ???????? ??????????? ???????????? ?? >>> ?????? https://people.ucsc.edu/~warner/buffer.html ? ???????? >>> ????????? 48x10/25 + 6-8/100 ?????????????? ???????? ?????? 32-40MB - >>> ????? ??????????. ? ???? ??????, ??? ???????????? ?????? - ??? 1/10G >>> ????????? (???? ?? 25), ?? ????? ????????????, ??? ???????? is >>> absolutely safe ? ??????? ?????? ??????? ? ????? ????????????. >>> >>> === 2. ????????????? inline notification ???????? ???????? ???????? >>> ??????? === >>> >>>> The advantage of ECN is that the sender is not placed in the >>>> position of being informed of a congestion condition well after the >>>> condition has occurred. Explicit notification allows the sender to >>>> be informed of a condition as it is forming so that it can take >>>> action while there is still a coherent ack pacing signal coming back >>>> from the receiver (that is before packet loss occurs). >>>> >>>> However, ECN is only a single bit marking. Is this enough? [ ... ] >>>> The conclusion from one presentation is that the single-bit marking, >>>> while coarse and non-specific is probably sufficient to moderate >>>> self-clocking TCP flows such that they do not place pressure on >>>> network buffers, leaving the buffers to deal with short term bursts >>>> from unconstrained sources. >>>> >>> [ ... ] >>>> >>>> And if we want to reduce buffer size and maintain efficient and fair >>>> performance how can we achieve it? One view is that sender pacing >>>> can remove much of the pressure on buffers, and self-clocking flows >>>> can stabilise without emitting transient bursts that need to be >>>> absorbed by buffers. Another view, and one that does not necessarily >>>> contradict the first, is that the self-clocking algorithm can >>>> operate with higher precision if there was some form of feedback >>>> from the network on the state of the network path. This can be as >>>> simple as a single bit (ECN) or a complete trace of path element >>>> queue state (HPCC). >>>> >>> >>> ??? ?? ????? ????? ?????? ????????? ECN >>> >>> === 3. flow-aware traffic management ?? ???????? ??????????? ??????? >>> ?? ????? === >>> >>>> If the network was in a position to be able to classify all >>>> currently active flows into either elephants or mice, then the >>>> network could be able to use different queuing regimes for each >>>> traffic class. This sorting adds to the cost and complexity of >>>> packet switches, and if scaling pressures are a factor in switch >>>> design then it?s not clear that the additional cost of switch >>>> complexity would be offset by a far superior efficiency outcome in >>>> the switching function. >>> >>> === 4. ? ???? ????? ??????? ???????? ???????? ??? ?????????? (???? >>> 100ms) ?????? === >>> >>>> Further analysis reveals an estimate of packet drop rates if the >>>> network?s buffers were reduced in size, and for this particular >>>> case, the analysis revealed that an 18msec buffer would be able to >>>> sustain a packet drop rate of less than 0.005%. >>> >>> ???????. >>> >>> >>> On 15.10.2021 14:48, Volodymyr Litovka wrote: >>>> >>>> ??????, >>>> >>>> ? ??? ??? ???? ????????, ? ? ?? ???? ?? ?????????? :) ? ???????, ?? >>>> ??????? ???? ??????? ?????? ?????, ??? ??? ?? ???? - >>>> >>>> ?? ????? ????????, ??????? ?????? (?? ???? ??? shared) ????????? ??? >>>> ????????? ??? ???????? ?? ??????????? ????? ????? ??????? ??? ????? >>>> traffic bursts? ?? ????? ????????? ?? ?????????? ?????????? ????? >>>> ?????????? ?? ??????? ?? ?????? "? ???? ??? ???????", "? ??????? >>>> ??????? ????? ? ????????? ?? ???????????" ?? ???? ????, ? ?? >>>> ?????????-??????????? - ???????? ?????? ? ????? ?????? ? ???????? >>>> ?????? ????????? ?????????? ??, ?????????, ?? ??????????? ?????????? >>>> 24x1G + 2x10G ? ???????????? ????????????????????? ??????????? >>>> >>>> ????? ??????? ??? ???????? ????????? :) >>>> >>>> ?????. >>>> >>>> -- >>>> Volodymyr Litovka >>>> ?? "Vision without Execution is Hallucination." -- Thomas Edison >>> >>> -- >>> Volodymyr Litovka >>> ?? "Vision without Execution is Hallucination." -- Thomas Edison >>> >>> >>> _______________________________________________ >>> uanog mailing list >>> uanog at uanog.kiev.ua >>> https://mailman.uanog.kiev.ua/mailman/listinfo/uanog >> _______________________________________________ >> uanog mailing list >> uanog at uanog.kiev.ua >> https://mailman.uanog.kiev.ua/mailman/listinfo/uanog > > -- > Volodymyr Litovka > "Vision without Execution is Hallucination." -- Thomas Edison > From doka at funlab.cc Tue Dec 28 12:11:56 2021 From: doka at funlab.cc (Volodymyr Litovka) Date: Tue, 28 Dec 2021 12:11:56 +0200 Subject: [uanog] switch's port buffers In-Reply-To: <5778237a-43b1-1c37-799c-1ad0079973e6@netassist.kiev.ua> References: <7c8d121e-64a9-9be0-5ba3-c9a2c3632f75@netassist.kiev.ua> <579150fd-e9c1-90e5-cfce-89d842dbc761@funlab.cc> <5778237a-43b1-1c37-799c-1ad0079973e6@netassist.kiev.ua> Message-ID: <18a3ef7a-d2f8-8dfc-8201-4450f1bcf5b3@funlab.cc> On 28.12.2021 12:00, Max Tulyev wrote: > ??? ????????, ?????? pause ? ????????????? ??? ????, ????? ?????? > ????? ? ????????? ?????????? ????! ?? ??? ?? ????????. ??????, ? > ?????, ??? ??? ???? ???-?? ??? ?????????, ?? ???? ??? ???? ? ???? ??????????? ? ?????????? ?????????? ????????? ? ????? ????????? ????. ? ???? ??????????? - Netflix, ? ?????????? - ? ?????????? > ? ??? ????? ????? ??? ????? ? ???? TCP - ??? ?? ???, ???????? ??????????. NAT ??? ???? ????? ? ????????? - ? ???? :) > > 28.12.21 11:36, Volodymyr Litovka ????: >> ??????, >> >> ???? ????????????, ??? ??? ??-?? ?????????? ???????? ????? ?? ??????? >> IP. ??????????? ?????????? ???????? ??????, ?????? ?? ???? ? >> ????????? underlay ????. ?????????? ? ?????? ECN - ??? ??????? IP ? >> ?????????? ?????????? ???????????? ?????? IP-?????????, ??????????? >> ????? ??????? ???????? ????? ?? ??????? ????. ???????? ??? ???????? >> ????? ???????? ? ???????????????? ? ??????????????? ????????????. >> >> On 28.12.2021 11:26, Max Tulyev wrote: >>> ??????! >>> >>> ????? ?????????????, ???????! >>> >>> ?????? ???? ??????: ?????? ?? ???????? ethernet flow control >>> ?????????? ????? ?????????? ????? ????? ????? ???????? ? ?????? >>> ???????. >>> >>> 28.12.21 11:01, Volodymyr Litovka ????: >>>> ??????, >>>> >>>> ??? ????????? ?????????, ???? ?????????? ?????? ??? ????????? >>>> ??????????? ? ????????? ???????? ?? ??????????? ???????? ? >>>> ???????????? ???????????? ? 2019 ???? - "*It hosted 98 attendees >>>> from 12 economies with 26 from academia and 72 from industry.*" -- >>>> https://blog.apnic.net/2019/12/12/sizing-the-buffer/ >>>> >>>> ??? ?????? ?? ????????? ?????? - ????????? ????? ???????? ???????? >>>> ???: http://buffer-workshop.stanford.edu/program/ >>>> >>>>> [ ... ] buffers also add additional lag to a packet?s transit >>>>> through the network. If you want to implement a low jitter >>>>> service, then deep buffers are decidedly unfriendly! The result is >>>>> the rather enigmatic observation that network buffers have to be >>>>> as big as they need to be, but no bigger! >>>> >>>> ??? ?????? ??? :) ??????? ?????? ?? ??????: >>>> >>>> === 1. ??????????? "????? 40MB - ??????" ?? ??????????? ???????? >>>> ?????????????? ?????????????? === >>>> >>>>> A study by a Stanford TCP research group in 2004 >>>>> used the central >>>>> limit theorem to point to a radically smaller model of buffer >>>>> size. Link efficiency can be maintained for N desynchronized flows >>>>> with a buffer that is dimensioned to the size of: >>>>> >>>>> *Size*?= (*/BW/*?? */RTT/*) / ?*/N/* >>>>> >>>>> This is a radical result for high-speed extended latency links in >>>>> a busy network. The consequences on router design are enormous: >>>>> >>>>> ??? ?For example, a 1 Tb/s ISP router carrying one TCP flow with an >>>>> ??? RTTmin of 100ms would require 12.5 GB of buffer and off-chip >>>>> ??? buffering. *If it carries 100,000 flows, then the buffer can be >>>>> ??? safely reduced to less than 40MB, reducing the buffering and >>>>> ??? worst-case latency by 99.7%*. With small buffers, the buffer >>>>> would >>>>> ??? comfortably fit on a single chip switch ASIC.? >>>>> >>>>> ??? Nick McKeown et. al. Sizing Router Buffers (Redux) >>>>> >>>>> >>>>> >>>> ????? ???????, ?????????? ?? ???????? ??????????? ???????????? ?? >>>> ?????? https://people.ucsc.edu/~warner/buffer.html ? ???????? >>>> ????????? 48x10/25 + 6-8/100 ?????????????? ???????? ?????? 32-40MB >>>> - ????? ??????????. ? ???? ??????, ??? ???????????? ?????? - ??? >>>> 1/10G ????????? (???? ?? 25), ?? ????? ????????????, ??? ???????? >>>> is absolutely safe ? ??????? ?????? ??????? ? ????? ????????????. >>>> >>>> === 2. ????????????? inline notification ???????? ???????? ???????? >>>> ??????? === >>>> >>>>> The advantage of ECN is that the sender is not placed in the >>>>> position of being informed of a congestion condition well after >>>>> the condition has occurred. Explicit notification allows the >>>>> sender to be informed of a condition as it is forming so that it >>>>> can take action while there is still a coherent ack pacing signal >>>>> coming back from the receiver (that is before packet loss occurs). >>>>> >>>>> However, ECN is only a single bit marking. Is this enough? [ ... ] >>>>> The conclusion from one presentation is that the single-bit >>>>> marking, while coarse and non-specific is probably sufficient to >>>>> moderate self-clocking TCP flows such that they do not place >>>>> pressure on network buffers, leaving the buffers to deal with >>>>> short term bursts from unconstrained sources. >>>>> >>>> [ ... ] >>>>> >>>>> And if we want to reduce buffer size and maintain efficient and >>>>> fair performance how can we achieve it? One view is that sender >>>>> pacing can remove much of the pressure on buffers, and >>>>> self-clocking flows can stabilise without emitting transient >>>>> bursts that need to be absorbed by buffers. Another view, and one >>>>> that does not necessarily contradict the first, is that the >>>>> self-clocking algorithm can operate with higher precision if there >>>>> was some form of feedback from the network on the state of the >>>>> network path. This can be as simple as a single bit (ECN) or a >>>>> complete trace of path element queue state (HPCC). >>>>> >>>> >>>> ??? ?? ????? ????? ?????? ????????? ECN >>>> >>>> === 3. flow-aware traffic management ?? ???????? ??????????? >>>> ??????? ?? ????? === >>>> >>>>> If the network was in a position to be able to classify all >>>>> currently active flows into either elephants or mice, then the >>>>> network could be able to use different queuing regimes for each >>>>> traffic class. This sorting adds to the cost and complexity of >>>>> packet switches, and if scaling pressures are a factor in switch >>>>> design then it?s not clear that the additional cost of switch >>>>> complexity would be offset by a far superior efficiency outcome in >>>>> the switching function. >>>> >>>> === 4. ? ???? ????? ??????? ???????? ???????? ??? ?????????? (???? >>>> 100ms) ?????? === >>>> >>>>> Further analysis reveals an estimate of packet drop rates if the >>>>> network?s buffers were reduced in size, and for this particular >>>>> case, the analysis revealed that an 18msec buffer would be able to >>>>> sustain a packet drop rate of less than 0.005%. >>>> >>>> ???????. >>>> >>>> >>>> On 15.10.2021 14:48, Volodymyr Litovka wrote: >>>>> >>>>> ??????, >>>>> >>>>> ? ??? ??? ???? ????????, ? ? ?? ???? ?? ?????????? :) ? ???????, >>>>> ?? ??????? ???? ??????? ?????? ?????, ??? ??? ?? ???? - >>>>> >>>>> ?? ????? ????????, ??????? ?????? (?? ???? ??? shared) ????????? >>>>> ??? ????????? ??? ???????? ?? ??????????? ????? ????? ??????? ??? >>>>> ????? traffic bursts? ?? ????? ????????? ?? ?????????? ?????????? >>>>> ????? ?????????? ?? ??????? ?? ?????? "? ???? ??? ???????", "? >>>>> ??????? ??????? ????? ? ????????? ?? ???????????" ?? ???? ????, ? >>>>> ?? ?????????-??????????? - ???????? ?????? ? ????? ?????? ? >>>>> ???????? ?????? ????????? ?????????? ??, ?????????, ?? ??????????? >>>>> ?????????? 24x1G + 2x10G ? ???????????? ????????????????????? >>>>> ??????????? >>>>> >>>>> ????? ??????? ??? ???????? ????????? :) >>>>> >>>>> ?????. >>>>> >>>>> -- >>>>> Volodymyr Litovka >>>>> ?? "Vision without Execution is Hallucination." -- Thomas Edison >>>> >>>> -- >>>> Volodymyr Litovka >>>> ?? "Vision without Execution is Hallucination." -- Thomas Edison >>>> >>>> >>>> _______________________________________________ >>>> uanog mailing list >>>> uanog at uanog.kiev.ua >>>> https://mailman.uanog.kiev.ua/mailman/listinfo/uanog >>> _______________________________________________ >>> uanog mailing list >>> uanog at uanog.kiev.ua >>> https://mailman.uanog.kiev.ua/mailman/listinfo/uanog >> >> -- >> Volodymyr Litovka >> ?? "Vision without Execution is Hallucination." -- Thomas Edison >> > _______________________________________________ > uanog mailing list > uanog at uanog.kiev.ua > https://mailman.uanog.kiev.ua/mailman/listinfo/uanog -- Volodymyr Litovka "Vision without Execution is Hallucination." -- Thomas Edison -------------- next part -------------- An HTML attachment was scrubbed... URL: From maxtul at netassist.kiev.ua Tue Dec 28 12:56:10 2021 From: maxtul at netassist.kiev.ua (Max Tulyev) Date: Tue, 28 Dec 2021 12:56:10 +0200 Subject: [uanog] switch's port buffers In-Reply-To: <18a3ef7a-d2f8-8dfc-8201-4450f1bcf5b3@funlab.cc> References: <7c8d121e-64a9-9be0-5ba3-c9a2c3632f75@netassist.kiev.ua> <579150fd-e9c1-90e5-cfce-89d842dbc761@funlab.cc> <5778237a-43b1-1c37-799c-1ad0079973e6@netassist.kiev.ua> <18a3ef7a-d2f8-8dfc-8201-4450f1bcf5b3@funlab.cc> Message-ID: 28.12.21 12:11, Volodymyr Litovka ????: > > On 28.12.2021 12:00, Max Tulyev wrote: >> ??? ????????, ?????? pause ? ????????????? ??? ????, ????? ?????? >> ????? ? ????????? ?????????? ????! ?? ??? ?? ????????. ??????, ? >> ?????, ??? ??? ???? ???-?? ??? ?????????, ?? ???? > > ??? ???? ? ???? ??????????? ? ?????????? ?????????? ????????? ? ????? > ????????? ????. ? ???? ??????????? - Netflix, ? ?????????? - ? ?????????? ???? ? ????????? ?????, ? ???? ? ???? ???? pause frames, ??? ??? ??????? ?? ????????? ?? ????????? ?? L2, ???? ??????? ???????. ??? ???? >> ? ??? ????? ????? ??? ????? ? ???? TCP - ??? ?? ???, ???????? ??????????. > > NAT ??? ???? ????? ? ????????? - ? ???? :) ? ???? ??? ?? ???????! :) From doka at funlab.cc Tue Dec 28 13:20:20 2021 From: doka at funlab.cc (Volodymyr Litovka) Date: Tue, 28 Dec 2021 13:20:20 +0200 Subject: [uanog] switch's port buffers In-Reply-To: References: <7c8d121e-64a9-9be0-5ba3-c9a2c3632f75@netassist.kiev.ua> <579150fd-e9c1-90e5-cfce-89d842dbc761@funlab.cc> <5778237a-43b1-1c37-799c-1ad0079973e6@netassist.kiev.ua> <18a3ef7a-d2f8-8dfc-8201-4450f1bcf5b3@funlab.cc> Message-ID: On 28.12.2021 12:56, Max Tulyev wrote: > >>> ??? ????????, ?????? pause ? ????????????? ??? ????, ????? ?????? >>> ????? ? ????????? ?????????? ????! ?? ??? ?? ????????. ??????, ? >>> ?????, ??? ??? ???? ???-?? ??? ?????????, ?? ???? >> >> ??? ???? ? ???? ??????????? ? ?????????? ?????????? ????????? ? ????? >> ????????? ????. ? ???? ??????????? - Netflix, ? ?????????? - ? >> ?????????? > > ???? ? ????????? ?????, ? ???? ? ???? ???? pause frames, ??? ??? > ??????? ?? ????????? ?? ????????? ?? L2, ???? ??????? ???????. ??? ???? ???? ????? ?????????? ? ?????????? ??? L2 :) >>> ? ??? ????? ????? ??? ????? ? ???? TCP - ??? ?? ???, ???????? >>> ??????????. >> >> NAT ??? ???? ????? ? ????????? - ? ???? :) > > ? ???? ??? ?? ???????! :) ?? ???? ???. ???? ????????????? known issues, ??????????? - ??? ? ????? ????????, ???? ??????? NAT ?? ?????? ???? :) ??? ?????, ??? ECN - ??? ???????????? ?????? ? ?????????? ?????????? ? ??????????????? ?????? ??? ????? ????? ? ?????????. ?? ???? ??? ?? "????? ? ?????????", ? "??????? ??, ??? ????????????? ? ????????? ???????????". -- Volodymyr Litovka "Vision without Execution is Hallucination." -- Thomas Edison -------------- next part -------------- An HTML attachment was scrubbed... URL: From alexander.moskalenko at gmail.com Tue Dec 28 18:40:29 2021 From: alexander.moskalenko at gmail.com (Oleksandr Moskalenko) Date: Tue, 28 Dec 2021 17:40:29 +0100 Subject: [uanog] =?utf-8?b?0JrQsNC90LDQuyDQtNC+INCk0YDQsNC90LrRhNGD0YA=?= =?utf-8?b?0YLQsA==?= Message-ID: ???????????, ??? ??????? ???????????? ??? ????? ???????????? ?????? ??????? 2 ?????? ????? ?????? (??????? 9 ? ?? ?? ????) ?? AWS Cage ?? ?????????? (Equinix FR5, Frankfurt, Germany, Interxion FRA6, Frankfurt, Germany? ????? 2 ?????? ??????? ?????????? ? ?????????????? ????. ???? ?????????? ?? ?????? ??? ???????? -------------- next part -------------- An HTML attachment was scrubbed... URL: From alexander.moskalenko at gmail.com Tue Dec 28 18:45:11 2021 From: alexander.moskalenko at gmail.com (Oleksandr Moskalenko) Date: Tue, 28 Dec 2021 17:45:11 +0100 Subject: [uanog] =?utf-8?b?0JrQsNC90LDQuyDQtNC+INCk0YDQsNC90LrRhNGD0YA=?= =?utf-8?b?0YLQsA==?= In-Reply-To: References: Message-ID: ????? ????????, ??? ?????? ?? 10???? + macsec ? ??? ?????? ???? ???? ????????? On Tue, Dec 28, 2021 at 5:40 PM Oleksandr Moskalenko < alexander.moskalenko at gmail.com> wrote: > ???????????, > > ??? ??????? ???????????? ??? ????? ???????????? ?????? ??????? 2 ?????? > ????? ?????? (??????? 9 ? ?? ?? ????) ?? AWS Cage ?? ?????????? (Equinix > FR5, Frankfurt, Germany, Interxion FRA6, Frankfurt, Germany? > ????? 2 ?????? ??????? ?????????? ? ?????????????? ????. > > ???? ?????????? ?? ?????? ??? ???????? > -------------- next part -------------- An HTML attachment was scrubbed... URL: From maxtul at netassist.kiev.ua Tue Dec 28 20:03:25 2021 From: maxtul at netassist.kiev.ua (Max Tulyev) Date: Tue, 28 Dec 2021 20:03:25 +0200 Subject: [uanog] =?utf-8?b?0JrQsNC90LDQuyDQtNC+INCk0YDQsNC90LrRhNGD0YA=?= =?utf-8?b?0YLQsA==?= In-Reply-To: References: Message-ID: <661fbe4d-0651-29b7-757a-cc98bcffb7e3@netassist.kiev.ua> ??? ???????? - ? ?????????? ????????? ))) ?????? ?? ? ?????? ??????, ? ?? ???????????? ????????, ????? ???? 28.12.21 18:45, Oleksandr Moskalenko ????: > ????? ????????, ??? ?????? ?? 10???? + macsec ? ??? ?????? ???? ???? > ????????? > > On Tue, Dec 28, 2021 at 5:40 PM Oleksandr Moskalenko > > > wrote: > > ???????????, > > ??? ??????? ???????????? ??? ????? ???????????? ?????? ??????? 2 > ?????? ????? ?????? (??????? 9 ? ?? ?? ????) ?? AWS Cage ?? > ?????????? (Equinix FR5, Frankfurt, Germany, Interxion FRA6, > Frankfurt, Germany? > ????? 2 ?????? ??????? ?????????? ? ?????????????? ????. > > ???? ?????????? ?? ?????? ??? ???????? > > > _______________________________________________ > uanog mailing list > uanog at uanog.kiev.ua > https://mailman.uanog.kiev.ua/mailman/listinfo/uanog From vladimir.sharun at ukr.net Tue Dec 28 22:48:01 2021 From: vladimir.sharun at ukr.net (Vladimir Sharun) Date: Tue, 28 Dec 2021 22:48:01 +0200 Subject: [uanog] =?utf-8?b?0JrQsNC90LDQuyDQtNC+INCk0YDQsNC90LrRhNGD0YA=?= =?utf-8?b?0YLQsA==?= In-Reply-To: <661fbe4d-0651-29b7-757a-cc98bcffb7e3@netassist.kiev.ua> References: <661fbe4d-0651-29b7-757a-cc98bcffb7e3@netassist.kiev.ua> Message-ID: <1640724175.721205000.veo3gt1z@frv53.fwdcdn.com> ??????, ???????? ??????? ? ???? ??? ?????????? DF (??? ????????????) ??? DWDM, ? ??? ??????, ??? L2 ??? VPN'? ?? ?????? (??? ?????????). macsec ?????? ????? ?? AES192, ??? ? VPN'?. ?? ? ?? DWDM, ?? ???????????, ???? ????????, ?? ???? ?? ???-?????? ????? ????????? ? ????????? ??????, ? ?????? ??????????? ???????? ?????. ??? ? ??? ??????????, ??? ??????? SFP-????? ??? ?????????????(?) ?? ?????? ?????? ????????. 28 ?????? 2021, 20:03:36, ??? "Max Tulyev" : ??? ???????? - ? ?????????? ????????? ))) ?????? ?? ? ?????? ??????, ? ?? ???????????? ????????, ????? ???? 28.12.21 18:45, Oleksandr Moskalenko ????: > ????? ????????, ??? ?????? ?? 10???? + macsec ? ??? ?????? ???? ???? > ????????? > > On Tue, Dec 28, 2021 at 5:40 PM Oleksandr Moskalenko > > > wrote: > > ???????????, > > ??? ??????? ???????????? ??? ????? ???????????? ?????? ??????? 2 > ?????? ????? ?????? (??????? 9 ? ?? ?? ????) ?? AWS Cage ?? > ?????????? (Equinix FR5, Frankfurt, Germany, Interxion FRA6, > Frankfurt, Germany? > ????? 2 ?????? ??????? ?????????? ? ?????????????? ????. > > ???? ?????????? ?? ?????? ??? ???????? > > > _______________________________________________ > uanog mailing list > uanog at uanog.kiev.ua > https://mailman.uanog.kiev.ua/mailman/listinfo/uanog _______________________________________________ uanog mailing list uanog at uanog.kiev.ua https://mailman.uanog.kiev.ua/mailman/listinfo/uanog -------------- next part -------------- An HTML attachment was scrubbed... URL: From alexander.moskalenko at gmail.com Tue Dec 28 23:02:36 2021 From: alexander.moskalenko at gmail.com (Oleksandr Moskalenko) Date: Tue, 28 Dec 2021 22:02:36 +0100 Subject: [uanog] =?utf-8?b?0JrQsNC90LDQuyDQtNC+INCk0YDQsNC90LrRhNGD0YA=?= =?utf-8?b?0YLQsA==?= In-Reply-To: <1640724175.721205000.veo3gt1z@frv53.fwdcdn.com> References: <661fbe4d-0651-29b7-757a-cc98bcffb7e3@netassist.kiev.ua> <1640724175.721205000.veo3gt1z@frv53.fwdcdn.com> Message-ID: ???? ?? ???????????? ??? ?????? ???????? ?? ?????? L2-emulation. On Tue, Dec 28, 2021 at 9:48 PM Vladimir Sharun wrote: > ??????, > > ???????? ??????? ? ???? ??? ?????????? DF (??? ????????????) ??? DWDM, ? > ??? ??????, ??? L2 ??? VPN'? ?? ?????? (??? ?????????). > > macsec ?????? ????? ?? AES192, ??? ? VPN'?. ?? ? ?? DWDM, ?? ???????????, > ???? ????????, ?? ???? ?? ???-?????? ????? ????????? ? ????????? ??????, ? > ?????? ??????????? ???????? ?????. ??? ? ??? ??????????, ??? ??????? > SFP-????? ??? ?????????????(?) ?? ?????? ?????? ????????. > > *28 ?????? 2021, 20:03:36, ??? "Max Tulyev" >: * > > ??? ???????? - ? ?????????? ????????? ))) > > ?????? ?? ? ?????? ??????, ? ?? ???????????? ????????, ????? ???? > > 28.12.21 18:45, Oleksandr Moskalenko ????: > > ????? ????????, ??? ?????? ?? 10???? + macsec ? ??? ?????? ???? ???? > > ????????? > > > > On Tue, Dec 28, 2021 at 5:40 PM Oleksandr Moskalenko > > > > > wrote: > > > > ???????????, > > > > ??? ??????? ???????????? ??? ????? ???????????? ?????? ??????? 2 > > ?????? ????? ?????? (??????? 9 ? ?? ?? ????) ?? AWS Cage ?? > > ?????????? (Equinix FR5, Frankfurt, Germany, Interxion FRA6, > > Frankfurt, Germany? > > ????? 2 ?????? ??????? ?????????? ? ?????????????? ????. > > > > ???? ?????????? ?? ?????? ??? ???????? > > > > > > _______________________________________________ > > uanog mailing list > > uanog at uanog.kiev.ua > > https://mailman.uanog.kiev.ua/mailman/listinfo/uanog > _______________________________________________ > uanog mailing listuanog at uanog.kiev.uahttps://mailman.uanog.kiev.ua/mailman/listinfo/uanog > > _______________________________________________ > uanog mailing list > uanog at uanog.kiev.ua > https://mailman.uanog.kiev.ua/mailman/listinfo/uanog -------------- next part -------------- An HTML attachment was scrubbed... URL: From vovik at dumpty.org Wed Dec 29 08:30:24 2021 From: vovik at dumpty.org (Volodymyr Yakovenko) Date: Wed, 29 Dec 2021 07:30:24 +0100 Subject: [uanog] =?utf-8?b?0JrQsNC90LDQuyDQtNC+INCk0YDQsNC90LrRhNGD0YA=?= =?utf-8?b?0YLQsA==?= In-Reply-To: References: <661fbe4d-0651-29b7-757a-cc98bcffb7e3@netassist.kiev.ua> <1640724175.721205000.veo3gt1z@frv53.fwdcdn.com> Message-ID: On Tue, Dec 28, 2021 at 10:02 PM Oleksandr Moskalenko < alexander.moskalenko at gmail.com> wrote: > ???? ?? ???????????? > ??? > ?????? ???????? ?? ?????? L2-emulation. > ????? - ??? ?????? MACsec ? MKA L2 transport ?????? ???????????? ????????? ??? ???? EtherType: - 0x888E - EAP over LAN (IEEE 802.1X) - 0x88E5 - MAC security (IEEE 802.1AE) DWDM wave ????? transparent ??? ???? EtherType. On Tue, Dec 28, 2021 at 9:48 PM Vladimir Sharun > wrote: > >> ??????, >> >> ???????? ??????? ? ???? ??? ?????????? DF (??? ????????????) ??? DWDM, ? >> ??? ??????, ??? L2 ??? VPN'? ?? ?????? (??? ?????????). >> >> macsec ?????? ????? ?? AES192, ??? ? VPN'?. ?? ? ?? DWDM, ?? ???????????, >> ???? ????????, ?? ???? ?? ???-?????? ????? ????????? ? ????????? ??????, ? >> ?????? ??????????? ???????? ?????. ??? ? ??? ??????????, ??? ??????? >> SFP-????? ??? ?????????????(?) ?? ?????? ?????? ????????. >> >> *28 ?????? 2021, 20:03:36, ??? "Max Tulyev" > >: * >> >> ??? ???????? - ? ?????????? ????????? ))) >> >> ?????? ?? ? ?????? ??????, ? ?? ???????????? ????????, ????? ???? >> >> 28.12.21 18:45, Oleksandr Moskalenko ????: >> > ????? ????????, ??? ?????? ?? 10???? + macsec ? ??? ?????? ???? ???? >> > ????????? >> > >> > On Tue, Dec 28, 2021 at 5:40 PM Oleksandr Moskalenko >> > > >> > wrote: >> > >> > ???????????, >> > >> > ??? ??????? ???????????? ??? ????? ???????????? ?????? ??????? 2 >> > ?????? ????? ?????? (??????? 9 ? ?? ?? ????) ?? AWS Cage ?? >> > ?????????? (Equinix FR5, Frankfurt, Germany, Interxion FRA6, >> > Frankfurt, Germany? >> > ????? 2 ?????? ??????? ?????????? ? ?????????????? ????. >> > >> > ???? ?????????? ?? ?????? ??? ???????? >> > >> > >> > _______________________________________________ >> > uanog mailing list >> > uanog at uanog.kiev.ua >> > https://mailman.uanog.kiev.ua/mailman/listinfo/uanog >> _______________________________________________ >> uanog mailing listuanog at uanog.kiev.uahttps://mailman.uanog.kiev.ua/mailman/listinfo/uanog >> >> _______________________________________________ >> uanog mailing list >> uanog at uanog.kiev.ua >> https://mailman.uanog.kiev.ua/mailman/listinfo/uanog > > _______________________________________________ > uanog mailing list > uanog at uanog.kiev.ua > https://mailman.uanog.kiev.ua/mailman/listinfo/uanog -- Regards, Volodymyr. -------------- next part -------------- An HTML attachment was scrubbed... URL: From vovik at dumpty.org Thu Dec 30 12:02:08 2021 From: vovik at dumpty.org (Volodymyr Yakovenko) Date: Thu, 30 Dec 2021 11:02:08 +0100 Subject: [uanog] switch's port buffers In-Reply-To: <5778237a-43b1-1c37-799c-1ad0079973e6@netassist.kiev.ua> References: <7c8d121e-64a9-9be0-5ba3-c9a2c3632f75@netassist.kiev.ua> <579150fd-e9c1-90e5-cfce-89d842dbc761@funlab.cc> <5778237a-43b1-1c37-799c-1ad0079973e6@netassist.kiev.ua> Message-ID: On Tue, Dec 28, 2021 at 11:00 AM Max Tulyev wrote: > ??? ????????, ?????? pause ? ????????????? ??? ????, ????? ?????? ????? > ? ????????? ?????????? ????! ?? ??? ?? ????????. ??????, ? ?????, ??? > ??? ???? ???-?? ??? ?????????, ?? ???? > ? ????????? Linux--<10GE>---Switch ? ????? Switch--<1GE>--Client pause frames (AKA Ethernet Flow Control) ?? ??????? - ????????????? Ethernet Flow Control per switchport of physical P2P link (???????? Switch-Linux). ???? Switch ????? ?? ????? pause ? Linux ?? ?? congestion ?? ????? Switch--<1GE>--Client ????? ??? ?? ????????????? transmission ?? Linux ?? ???? IP destinations. ? ??????? ??????? EPL (Ethernet Private Line) ????????? ????????? ?????????? pause ? ?????????? ???? ? ??????, ????????, ???? EPL bitrate > contracted CIR. ? ??? ????? ????? ??? ????? ? ???? TCP - ??? ?? ???, ???????? ??????????. > ??? ? ?????? ? ECN ????? TCP ???? ?? ???????, ?? ???????????? ??? ????????? ???? Traffic Class ? IP header (??? ????? ? DSCP ? ?? ????????? ?????? in-transit). TCP client - ?? ????????? IP ?????? ? ECN ?????? ? CE (Congestion Encountered) ? - ??? ??????? ??? Server ? Client ??? TCP setup negotiated ECN supports ???????? TCP server ECN-Echo (ECE), ??????? ??????? Server ?????? ??? ?? TCP ?????? ? Client ? ???? congestion. > 28.12.21 11:36, Volodymyr Litovka ????: > > ??????, > > > > ???? ????????????, ??? ??? ??-?? ?????????? ???????? ????? ?? ??????? > > IP. ??????????? ?????????? ???????? ??????, ?????? ?? ???? ? ????????? > > underlay ????. ?????????? ? ?????? ECN - ??? ??????? IP ? ?????????? > > ?????????? ???????????? ?????? IP-?????????, ??????????? ????? ??????? > > ???????? ????? ?? ??????? ????. ???????? ??? ???????? ????? ???????? ? > > ???????????????? ? ??????????????? ????????????. > > > > On 28.12.2021 11:26, Max Tulyev wrote: > >> ??????! > >> > >> ????? ?????????????, ???????! > >> > >> ?????? ???? ??????: ?????? ?? ???????? ethernet flow control > >> ?????????? ????? ?????????? ????? ????? ????? ???????? ? ?????? ???????. > >> > >> 28.12.21 11:01, Volodymyr Litovka ????: > >>> ??????, > >>> > >>> ??? ????????? ?????????, ???? ?????????? ?????? ??? ????????? > >>> ??????????? ? ????????? ???????? ?? ??????????? ???????? ? > >>> ???????????? ???????????? ? 2019 ???? - "*It hosted 98 attendees from > >>> 12 economies with 26 from academia and 72 from industry.*" -- > >>> https://blog.apnic.net/2019/12/12/sizing-the-buffer/ > >>> > >>> ??? ?????? ?? ????????? ?????? - ????????? ????? ???????? ???????? > >>> ???: http://buffer-workshop.stanford.edu/program/ > >>> > >>>> [ ... ] buffers also add additional lag to a packet?s transit > >>>> through the network. If you want to implement a low jitter service, > >>>> then deep buffers are decidedly unfriendly! The result is the rather > >>>> enigmatic observation that network buffers have to be as big as they > >>>> need to be, but no bigger! > >>> > >>> ??? ?????? ??? :) ??????? ?????? ?? ??????: > >>> > >>> === 1. ??????????? "????? 40MB - ??????" ?? ??????????? ???????? > >>> ?????????????? ?????????????? === > >>> > >>>> A study by a Stanford TCP research group in 2004 > >>>> used the central limit > >>>> theorem to point to a radically smaller model of buffer size. Link > >>>> efficiency can be maintained for N desynchronized flows with a > >>>> buffer that is dimensioned to the size of: > >>>> > >>>> *Size* = (*/BW/* ? */RTT/*) / ?*/N/* > >>>> > >>>> This is a radical result for high-speed extended latency links in a > >>>> busy network. The consequences on router design are enormous: > >>>> > >>>> ?For example, a 1 Tb/s ISP router carrying one TCP flow with an > >>>> RTTmin of 100ms would require 12.5 GB of buffer and off-chip > >>>> buffering. *If it carries 100,000 flows, then the buffer can be > >>>> safely reduced to less than 40MB, reducing the buffering and > >>>> worst-case latency by 99.7%*. With small buffers, the buffer would > >>>> comfortably fit on a single chip switch ASIC.? > >>>> > >>>> Nick McKeown et. al. Sizing Router Buffers (Redux) > >>>> < > https://ccronline.sigcomm.org/2019/ccr-october-2019/sizing-router-buffers-redux/ > > > >>>> > >>> ????? ???????, ?????????? ?? ???????? ??????????? ???????????? ?? > >>> ?????? https://people.ucsc.edu/~warner/buffer.html ? ???????? > >>> ????????? 48x10/25 + 6-8/100 ?????????????? ???????? ?????? 32-40MB - > >>> ????? ??????????. ? ???? ??????, ??? ???????????? ?????? - ??? 1/10G > >>> ????????? (???? ?? 25), ?? ????? ????????????, ??? ???????? is > >>> absolutely safe ? ??????? ?????? ??????? ? ????? ????????????. > >>> > >>> === 2. ????????????? inline notification ???????? ???????? ???????? > >>> ??????? === > >>> > >>>> The advantage of ECN is that the sender is not placed in the > >>>> position of being informed of a congestion condition well after the > >>>> condition has occurred. Explicit notification allows the sender to > >>>> be informed of a condition as it is forming so that it can take > >>>> action while there is still a coherent ack pacing signal coming back > >>>> from the receiver (that is before packet loss occurs). > >>>> > >>>> However, ECN is only a single bit marking. Is this enough? [ ... ] > >>>> The conclusion from one presentation is that the single-bit marking, > >>>> while coarse and non-specific is probably sufficient to moderate > >>>> self-clocking TCP flows such that they do not place pressure on > >>>> network buffers, leaving the buffers to deal with short term bursts > >>>> from unconstrained sources. > >>>> > >>> [ ... ] > >>>> > >>>> And if we want to reduce buffer size and maintain efficient and fair > >>>> performance how can we achieve it? One view is that sender pacing > >>>> can remove much of the pressure on buffers, and self-clocking flows > >>>> can stabilise without emitting transient bursts that need to be > >>>> absorbed by buffers. Another view, and one that does not necessarily > >>>> contradict the first, is that the self-clocking algorithm can > >>>> operate with higher precision if there was some form of feedback > >>>> from the network on the state of the network path. This can be as > >>>> simple as a single bit (ECN) or a complete trace of path element > >>>> queue state (HPCC). > >>>> > >>> > >>> ??? ?? ????? ????? ?????? ????????? ECN > >>> > >>> === 3. flow-aware traffic management ?? ???????? ??????????? ??????? > >>> ?? ????? === > >>> > >>>> If the network was in a position to be able to classify all > >>>> currently active flows into either elephants or mice, then the > >>>> network could be able to use different queuing regimes for each > >>>> traffic class. This sorting adds to the cost and complexity of > >>>> packet switches, and if scaling pressures are a factor in switch > >>>> design then it?s not clear that the additional cost of switch > >>>> complexity would be offset by a far superior efficiency outcome in > >>>> the switching function. > >>> > >>> === 4. ? ???? ????? ??????? ???????? ???????? ??? ?????????? (???? > >>> 100ms) ?????? === > >>> > >>>> Further analysis reveals an estimate of packet drop rates if the > >>>> network?s buffers were reduced in size, and for this particular > >>>> case, the analysis revealed that an 18msec buffer would be able to > >>>> sustain a packet drop rate of less than 0.005%. > >>> > >>> ???????. > >>> > >>> > >>> On 15.10.2021 14:48, Volodymyr Litovka wrote: > >>>> > >>>> ??????, > >>>> > >>>> ? ??? ??? ???? ????????, ? ? ?? ???? ?? ?????????? :) ? ???????, ?? > >>>> ??????? ???? ??????? ?????? ?????, ??? ??? ?? ???? - > >>>> > >>>> ?? ????? ????????, ??????? ?????? (?? ???? ??? shared) ????????? ??? > >>>> ????????? ??? ???????? ?? ??????????? ????? ????? ??????? ??? ????? > >>>> traffic bursts? ?? ????? ????????? ?? ?????????? ?????????? ????? > >>>> ?????????? ?? ??????? ?? ?????? "? ???? ??? ???????", "? ??????? > >>>> ??????? ????? ? ????????? ?? ???????????" ?? ???? ????, ? ?? > >>>> ?????????-??????????? - ???????? ?????? ? ????? ?????? ? ???????? > >>>> ?????? ????????? ?????????? ??, ?????????, ?? ??????????? ?????????? > >>>> 24x1G + 2x10G ? ???????????? ????????????????????? ??????????? > >>>> > >>>> ????? ??????? ??? ???????? ????????? :) > >>>> > >>>> ?????. > >>>> > >>>> -- > >>>> Volodymyr Litovka > >>>> "Vision without Execution is Hallucination." -- Thomas Edison > >>> > >>> -- > >>> Volodymyr Litovka > >>> "Vision without Execution is Hallucination." -- Thomas Edison > >>> > >>> > >>> _______________________________________________ > >>> uanog mailing list > >>> uanog at uanog.kiev.ua > >>> https://mailman.uanog.kiev.ua/mailman/listinfo/uanog > >> _______________________________________________ > >> uanog mailing list > >> uanog at uanog.kiev.ua > >> https://mailman.uanog.kiev.ua/mailman/listinfo/uanog > > > > -- > > Volodymyr Litovka > > "Vision without Execution is Hallucination." -- Thomas Edison > > > _______________________________________________ > uanog mailing list > uanog at uanog.kiev.ua > https://mailman.uanog.kiev.ua/mailman/listinfo/uanog -- Regards, Volodymyr. -------------- next part -------------- An HTML attachment was scrubbed... URL: From maksym at tulyuk.com Thu Dec 30 12:56:51 2021 From: maksym at tulyuk.com (Maksym Tulyuk) Date: Thu, 30 Dec 2021 11:56:51 +0100 Subject: [uanog] switch's port buffers In-Reply-To: References: Message-ID: <44e95a0e-d070-4f9a-9e7d-43152cf2522c@Spark> ??????! ? ???? ?????? ?????? ??????: 1) ?? ?????????? ??????? "We still know very little about buffer size? ? ???? ??? ?????? ?????/????? 2) ??? ????????? ?????????? ????????? TCP tuning ? ?? ????? ??????, ??? ??? ? 2019 ? Deutsche Telekom ?? ????????? ????? 84.4% (??? ????????? - UDP ? ??????? ?????? QUIC) 3) ??? ??????? ???????????? ??? MTU size 1500, ???? ? ???????? ???? ?????? ??????? 50% ???????? https://stats.ams-ix.net/sflow/size.html ? ??? ????? ?? ??????? 35% - ??? ?????? ???????? 64-127 bytes ?? ???????????: ? Understanding switch buffer utilization in CLOS data center fabric (Verizon) http://buffer-workshop.stanford.edu/slides/Understanding%20switch%20buffer%20utilization%20in%20CLOS%20data%20center%20fabric.pptx ? Buffer sizing and Video QoE Measurements at Netflix http://buffer-workshop.stanford.edu/slides/netflix%20buffer%20sizing.pdf ? Queueing at the Telco Service Edge (Deutsche Telekom) http://buffer-workshop.stanford.edu/slides/BS_QueueingEdge_2019_12_03.pdf ? Buffer Sizing Experiments at Facebook?http://buffer-workshop.stanford.edu/papers/paper30.pdf ?? ?????????: ??????? ????? ??????????, ????? ???????? ?????? ?? 2 ???????? http://buffer-workshop.stanford.edu/papers/paper18.pdf ?? ??? ??? ???? ? 2019, ? ?????? ??????? ??? ???? ??? ???? ? ??????? ??? UDP+QUIC ? ??? ???????????? ?? ???? ??????? ????? ???????? ? ????. ?????? On 28 Dec 2021, 10:02 +0100, Volodymyr Litovka , wrote: > ??????, > ??? ????????? ?????????, ???? ?????????? ?????? ??? ????????? ??????????? ? ????????? ???????? ?? ??????????? ???????? ? ???????????? ???????????? ? 2019 ???? - "It hosted 98 attendees from 12 economies with 26 from academia and 72 from industry." -- https://blog.apnic.net/2019/12/12/sizing-the-buffer/ > ??? ?????? ?? ????????? ?????? - ????????? ????? ???????? ???????? ???: http://buffer-workshop.stanford.edu/program/ > > [ ... ] buffers also add additional lag to a packet?s transit through the network. If you want to implement a low jitter service, then deep buffers are decidedly unfriendly! The result is the rather enigmatic observation that network buffers have to be as big as they need to be, but no bigger! > ??? ?????? ??? :) ??????? ?????? ?? ??????: > === 1. ??????????? "????? 40MB - ??????" ?? ??????????? ???????? ?????????????? ?????????????? === > > A study by a Stanford TCP research group in 2004 used the central limit theorem to point to a radically smaller model of buffer size. Link efficiency can be maintained for N desynchronized flows with a buffer that is dimensioned to the size of: > > Size?= (BW???RTT) / ?N > > This is a radical result for high-speed extended latency links in a busy network. The consequences on router design are enormous: > > > ?For example, a 1 Tb/s ISP router carrying one TCP flow with an RTTmin of 100ms would require 12.5 GB of buffer and off-chip buffering. If it carries 100,000 flows, then the buffer can be safely reduced to less than 40MB, reducing the buffering and worst-case latency by 99.7%. With small buffers, the buffer would comfortably fit on a single chip switch ASIC.? > > > Nick McKeown et. al. Sizing Router Buffers (Redux) > ????? ???????, ?????????? ?? ???????? ??????????? ???????????? ?? ?????? https://people.ucsc.edu/~warner/buffer.html ? ???????? ????????? 48x10/25 + 6-8/100 ?????????????? ???????? ?????? 32-40MB - ????? ??????????. ? ???? ??????, ??? ???????????? ?????? - ??? 1/10G ????????? (???? ?? 25), ?? ????? ????????????, ??? ???????? is absolutely safe ? ??????? ?????? ??????? ? ????? ????????????. > === 2. ????????????? inline notification ???????? ???????? ???????? ??????? === > > The advantage of ECN is that the sender is not placed in the position of being informed of a congestion condition well after the condition has occurred. Explicit notification allows the sender to be informed of a condition as it is forming so that it can take action while there is still a coherent ack pacing signal coming back from the receiver (that is before packet loss occurs). > > However, ECN is only a single bit marking. Is this enough? [ ... ] The conclusion from one presentation is that the single-bit marking, while coarse and non-specific is probably sufficient to moderate self-clocking TCP flows such that they do not place pressure on network buffers, leaving the buffers to deal with short term bursts from unconstrained sources. > [ ... ] > > And if we want to reduce buffer size and maintain efficient and fair performance how can we achieve it? One view is that sender pacing can remove much of the pressure on buffers, and self-clocking flows can stabilise without emitting transient bursts that need to be absorbed by buffers. Another view, and one that does not necessarily contradict the first, is that the self-clocking algorithm can operate with higher precision if there was some form of feedback from the network on the state of the network path. This can be as simple as a single bit (ECN) or a complete trace of path element queue state (HPCC). > > ??? ?? ????? ????? ?????? ????????? ECN > === 3. flow-aware traffic management ?? ???????? ??????????? ??????? ?? ????? === > > If the network was in a position to be able to classify all currently active flows into either elephants or mice, then the network could be able to use different queuing regimes for each traffic class. This sorting adds to the cost and complexity of packet switches, and if scaling pressures are a factor in switch design then it?s not clear that the additional cost of switch complexity would be offset by a far superior efficiency outcome in the switching function. > > === 4. ? ???? ????? ??????? ???????? ???????? ??? ?????????? (???? 100ms) ?????? === > > Further analysis reveals an estimate of packet drop rates if the network?s buffers were reduced in size, and for this particular case, the analysis revealed that an 18msec buffer would be able to sustain a packet drop rate of less than 0.005%. > > ???????. > > On 15.10.2021 14:48, Volodymyr Litovka wrote: > > ??????, > > ? ??? ??? ???? ????????, ? ? ?? ???? ?? ?????????? :) ? ???????, ?? ??????? ???? ??????? ?????? ?????, ??? ??? ?? ???? - > > ?? ????? ????????, ??????? ?????? (?? ???? ??? shared) ????????? ??? ????????? ??? ???????? ?? ??????????? ????? ????? ??????? ??? ????? traffic bursts? ?? ????? ????????? ?? ?????????? ?????????? ????? ?????????? ?? ??????? ?? ?????? "? ???? ??? ???????", "? ??????? ??????? ????? ? ????????? ?? ???????????" ?? ???? ????, ? ?? ?????????-??????????? - ???????? ?????? ? ????? ?????? ? ???????? ?????? ????????? ?????????? ??, ?????????, ?? ??????????? ?????????? 24x1G + 2x10G ? ???????????? ????????????????????? ??????????? > > ????? ??????? ??? ???????? ????????? :) > > ?????. > > -- > > Volodymyr Litovka > > "Vision without Execution is Hallucination." -- Thomas Edison > -- > Volodymyr Litovka > "Vision without Execution is Hallucination." -- Thomas Edison > _______________________________________________ > uanog mailing list > uanog at uanog.kiev.ua > https://mailman.uanog.kiev.ua/mailman/listinfo/uanog -------------- next part -------------- An HTML attachment was scrubbed... URL: From doka at funlab.cc Thu Dec 30 13:43:33 2021 From: doka at funlab.cc (Volodymyr Litovka) Date: Thu, 30 Dec 2021 13:43:33 +0200 Subject: [uanog] switch's port buffers In-Reply-To: <44e95a0e-d070-4f9a-9e7d-43152cf2522c@Spark> References: <44e95a0e-d070-4f9a-9e7d-43152cf2522c@Spark> Message-ID: On 30.12.2021 12:56, Maksym Tulyuk wrote: > ? ???? ?????? ?????? ??????: > [ ... ] > 2) ??? ????????? ?????????? ????????? TCP tuning ? ?? ????? ??????, > ??? ??? ? 2019 ? Deutsche Telekom ?? ????????? ????? 84.4% (??? > ????????? - UDP ? ??????? ?????? QUIC) ???, ???? ????????? * ????? underlaying UDP ??? QUIC ????????? ????????? ? ???? ?????? ???????????? flow control, ??????? ????? ??? ?? /_?????_/ ????????? ?? ECN bit ? IP-????????? ? ?????? ???????? ????? ? ?????? QUIC flow control * ??????? "flow" ????????? ??? ? TCP-??????? (????? ??? ?????????? ??????? ?? TCP session), ??? ? ? UDP-??????? (src/dst ip/port), ??????, ? ???????, ????? ????????? ??????? ??? "100,000 flows", ?? ???? ???? ?????? ?? ??????????/??????????? ??????? ??????, ? ?? ? TCP sessions > 3) ??? ??????? ???????????? ??? MTU size 1500, ???? ? ???????? ???? > ?????? ??????? 50% ???????? https://stats.ams-ix.net/sflow/size.html ? > ??? ????? ?? ??????? 35% - ??? ?????? ???????? 64-127 bytes ???????, ??? ??? ?? ????? ???????? ???????? ??? ????????? ??????????. ??? ???????? ???????????? ??????? 1500, ?? ??????? ?????????? ???????? in general ?? ???????????? ??????? ?? ???????? ??????????. > ?? ??? ??? ???? ? 2019, ? ?????? ??????? ??? ???? ??? ???? ? ??????? > ??? UDP+QUIC ? ??? ???????????? ?? ???? ??????? ????? ???????? ? ????. ? ?????? "???" ????, still ???, ?? ???????? - (1) QUIC ????? ???????????? ECN ? (2) ???????? ???????? (?????????? ? ?????????? ?????? ? ????????) ?? ?????????? ? ??????????? ?? underlying protocol - ????? ??? ??? ?? ????? ???????? ?? ???????? ???????, ? bigdata ??? ????? ????? ??????? ?? ????? ????????, ?????? ?????? deep buffers ??? ??????????????? ???????? :) ??????????, ??????????? ???????? ???????????? ? ???????? ???????????? ??? - ???? ???????? ????? ? ?????????? ? 40-80MB shared buffer, ? ???? - ????????? ? ??????? ?????? ???? - 40-80MB ? 4-8GB ????????. ? ??????????????? ????????: ??? ????????? ?????????? - ??????, ??? ????????? ??????? (??????) - ??????. ??????? -- Volodymyr Litovka "Vision without Execution is Hallucination." -- Thomas Edison -------------- next part -------------- An HTML attachment was scrubbed... URL: From violent at death.kiev.ua Thu Dec 30 14:06:55 2021 From: violent at death.kiev.ua (Igor Grabin) Date: Thu, 30 Dec 2021 14:06:55 +0200 Subject: [uanog] switch's port buffers In-Reply-To: References: <20211018155841.GY5646@death.kiev.ua> Message-ID: <20211230120655.GO30278@death.kiev.ua> On Mon, Oct 18, 2021 at 11:09:25PM +0300, Dmytro Cherkasov wrote: > ??, ?? ??? ?????? ?????????? ????????? ?? ???'??????? ??????? ????????? > ????????? ????????. ? ????????, ????? ????????? ????????? ????? ???? > ??????????? ????????, ? ???????????? ?????? ? ??? ?????? ????????? ???? 8 > ?????????? ????? ?? 8 ????????. ? ?? ?? ?? ??????? ? ??? ????????, ??? ??? > "?????"? ??-???????? - c4948-10GE. ??? ??????????? ???????, ?? ??????? ??????? ?? 10 ? ??????? ?? ?????????? ?????????? ?????-??????? "???? ????". ????? ??????? - dell n1548 ? ????, ???? c4948E. ???? ??? ?????? trident 2+ mbr, -- Igor "CacoDem0n" Grabin From maksym at tulyuk.com Thu Dec 30 21:53:46 2021 From: maksym at tulyuk.com (Maksym Tulyuk) Date: Thu, 30 Dec 2021 20:53:46 +0100 Subject: [uanog] switch's port buffers In-Reply-To: References: <44e95a0e-d070-4f9a-9e7d-43152cf2522c@Spark> Message-ID: ?? ??????? ?????: ???????? ? ???????????? ?????? ???? ???????????? ? ?? - ?????? ? ????????. ???????, ????? ?????? ? ???????? ????????, ? ??? ?????? ???????? QoS ;) P.S. ???????????? ? QUIC ??? ?????? ?? ???????????? ???????, ?? ???????????? ??????????? ?????? ????? we concluded that integrating QUIC in our apps would reduce the tail-end latencies compared to TCP. We witnessed a reduction of 10-30 percent in tail-end latencies for HTTPS traffic at scale in our rider and driver apps. In addition to improving the performance of our apps in low connectivity networks, QUIC gives us end-to-end control over the flow of packets in the user space (c) Uber devs Our tests have shown that QUIC offers improvements on several metrics. People on Facebook experienced a 6 percent reduction in request errors, a 20 percent tail latency reduction, and a 5 percent reduction in response header size relative to HTTP/2. This had cascading effects on other metrics as well, indicating that peoples? experience was greatly enhanced by QUIC (c) Facebook devs The experiments showed that QUIC had a transformative effect on video metrics in the Facebook app. Mean time between rebuffering (MTBR), a measure of the time between buffering events, improved in aggregate by up to 22 percent, depending on the platform. The overall error count on video requests was reduced by 8 percent. The rate of video stalls was reduced by 20 percent. (C) ) Facebook devs ? ??? ???? ????? ?? ?????? ?? ????????. ?????? On 30 Dec 2021, 12:43 +0100, Volodymyr Litovka , wrote: On 30.12.2021 12:56, Maksym Tulyuk wrote: ? ???? ?????? ?????? ??????: [ ... ] 2) ??? ????????? ?????????? ????????? TCP tuning ? ?? ????? ??????, ??? ??? ? 2019 ? Deutsche Telekom ?? ????????? ????? 84.4% (??? ????????? - UDP ? ??????? ?????? QUIC) ???, ???? ????????? ? ????? underlaying UDP ??? QUIC ????????? ????????? ? ???? ?????? ???????????? flow control, ??????? ????? ??? ?? ????? ????????? ?? ECN bit ? IP-????????? ? ?????? ???????? ????? ? ?????? QUIC flow control ? ??????? "flow" ????????? ??? ? TCP-??????? (????? ??? ?????????? ??????? ?? TCP session), ??? ? ? UDP-??????? (src/dst ip/port), ??????, ? ???????, ????? ????????? ??????? ??? "100,000 flows", ?? ???? ???? ?????? ?? ??????????/??????????? ??????? ??????, ? ?? ? TCP sessions 3) ??? ??????? ???????????? ??? MTU size 1500, ???? ? ???????? ???? ?????? ??????? 50% ???????? https://stats.ams-ix.net/sflow/size.html ? ??? ????? ?? ??????? 35% - ??? ?????? ???????? 64-127 bytes ???????, ??? ??? ?? ????? ???????? ???????? ??? ????????? ??????????. ??? ???????? ???????????? ??????? 1500, ?? ??????? ?????????? ???????? in general ?? ???????????? ??????? ?? ???????? ??????????. ?? ??? ??? ???? ? 2019, ? ?????? ??????? ??? ???? ??? ???? ? ??????? ??? UDP+QUIC ? ??? ???????????? ?? ???? ??????? ????? ???????? ? ????. ? ?????? "???" ????, still ???, ?? ???????? - (1) QUIC ????? ???????????? ECN ? (2) ???????? ???????? (?????????? ? ?????????? ?????? ? ????????) ?? ?????????? ? ??????????? ?? underlying protocol - ????? ??? ??? ?? ????? ???????? ?? ???????? ???????, ? bigdata ??? ????? ????? ??????? ?? ????? ????????, ?????? ?????? deep buffers ??? ??????????????? ???????? :) ??????????, ??????????? ???????? ???????????? ? ???????? ???????????? ??? - ???? ???????? ????? ? ?????????? ? 40-80MB shared buffer, ? ???? - ????????? ? ??????? ?????? ???? - 40-80MB ? 4-8GB ????????. ? ??????????????? ????????: ??? ????????? ?????????? - ??????, ??? ????????? ??????? (??????) - ??????. ??????? -- Volodymyr Litovka "Vision without Execution is Hallucination." -- Thomas Edison _______________________________________________ uanog mailing list uanog at uanog.kiev.ua https://mailman.uanog.kiev.ua/mailman/listinfo/uanog -------------- next part -------------- An HTML attachment was scrubbed... URL: From doka at funlab.cc Thu Dec 30 22:33:59 2021 From: doka at funlab.cc (Volodymyr Litovka) Date: Thu, 30 Dec 2021 22:33:59 +0200 Subject: [uanog] switch's port buffers In-Reply-To: References: <44e95a0e-d070-4f9a-9e7d-43152cf2522c@Spark> Message-ID: <2967b455-a1ae-0f79-a75e-60ffe469caed@funlab.cc> On 30.12.2021 21:53, Maksym Tulyuk wrote: > ?? ??????? ?????: ???????? ? ???????????? ?????? ???? ???????????? ? > ?? - ?????? ? ????????. ???????, ????? ?????? ? ???????? ????????, ? > ??? ?????? ???????? QoS ;) ???? ?? ??? ??????? ?? ??????? ??? ???????????? ?????? - ?? ? ???? ?????? ?????????? ??? ??? ?????? ???????????. ???? ?? ?????? ?????????? - ?? ??, ???????????? ????? ????? ??????????. > P.S. ???????????? ? QUIC ??? ?????? ?? ???????????? ???????, ?? > ???????????? ??????????? ?????? ????? [ ... ] ? ??? ???? ????? ?? > ?????? ?? ????????. ???????, ??? app layer ????? ?????? ?? ?????? ?? ???? ? ??????????? ???. ?? ???? app layer ????? ???????? ???????, ?????????? ?????, ?? ????????? ????? ???? ?????. -- Volodymyr Litovka "Vision without Execution is Hallucination." -- Thomas Edison -------------- next part -------------- An HTML attachment was scrubbed... URL: From doctorchd at gmail.com Thu Dec 30 22:53:52 2021 From: doctorchd at gmail.com (Dmytro Cherkasov) Date: Thu, 30 Dec 2021 22:53:52 +0200 Subject: [uanog] switch's port buffers In-Reply-To: <20211230120655.GO30278@death.kiev.ua> References: <20211018155841.GY5646@death.kiev.ua> <20211230120655.GO30278@death.kiev.ua> Message-ID: ??, ? ????????, ??? ? ????? Dell N1548. ????? ??????? ????-???? ????, ???? ? ????? ??????? ?????, ??? ????????????? ??????? ? ? ?????? ???????????. ?????, ?? ? ??? ???????? -- ???? ?????? ? 2?10Gbps ? 20 ????????, ???? ?? ???? ??????? ??? ??? ? ???? ??????????? ????????. IMHO ??? ?????? ???? ????????, ??? ?????? ???? ????? ??????? ??????. ????????? ?????????? ??????? ? ?????? ???????? ??? ?? ????, ????????? ?????? ????????? ????????? ??? ??????????? ???????. ???? ????? ??????? ? ????? ?????????? ??????? -- 100? ??????? ? ?????? 1 Tbps, ?? ???????? ? ?????????? ?? 10 Mbps ?? ?????, ?????? ???????? ???? ??? ??????????? ??????????. ??, 30 ????. 2021 ?. ? 14:06 Igor Grabin ????: > On Mon, Oct 18, 2021 at 11:09:25PM +0300, Dmytro Cherkasov wrote: > > ??, ?? ??? ?????? ?????????? ????????? ?? ???'??????? ??????? ????????? > > ????????? ????????. ? ????????, ????? ????????? ????????? ????? ???? > > ??????????? ????????, ? ???????????? ?????? ? ??? ?????? ????????? ???? 8 > > ?????????? ????? ?? 8 ????????. ? ?? ?? ?? ??????? ? ??? ????????, ??? > ??? > > "?????"? > ??-???????? - c4948-10GE. ??? ??????????? ???????, ?? ??????? ??????? > ?? 10 ? ??????? ?? ?????????? ?????????? ?????-??????? "???? ????". > > ????? ??????? - dell n1548 ? ????, ???? c4948E. ???? ??? ?????? > trident 2+ > > mbr, > -- > Igor "CacoDem0n" Grabin > -------------- next part -------------- An HTML attachment was scrubbed... URL: From violent at death.kiev.ua Fri Dec 31 10:32:45 2021 From: violent at death.kiev.ua (Igor Grabin) Date: Fri, 31 Dec 2021 10:32:45 +0200 Subject: [uanog] switch's port buffers In-Reply-To: References: <20211018155841.GY5646@death.kiev.ua> <20211230120655.GO30278@death.kiev.ua> Message-ID: <20211231083245.GP30278@death.kiev.ua> On Thu, Dec 30, 2021 at 10:53:52PM +0200, Dmytro Cherkasov wrote: > ??, ? ????????, ??? ? ????? Dell N1548. ????, ???????, ??? ? ??? ??????????. ???????? ??????? ?????? ??? ???? ???????? - N3048. ???? ??????????? ??? S-?????, ?? ???????, ??? os10 - ????? ;-) mbr, -- Igor "CacoDem0n" Grabin