idnits 2.17.1 draft-rfced-info-partridge-00.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** Cannot find the required boilerplate sections (Copyright, IPR, etc.) in this document. Expected boilerplate is as follows today (2024-04-23) according to https://trustee.ietf.org/license-info : IETF Trust Legal Provisions of 28-dec-2009, Section 6.a: This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79. IETF Trust Legal Provisions of 28-dec-2009, Section 6.b(i), paragraph 2: Copyright (c) 2024 IETF Trust and the persons identified as the document authors. All rights reserved. IETF Trust Legal Provisions of 28-dec-2009, Section 6.b(i), paragraph 3: This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- ** Missing expiration date. The document expiration date should appear on the first and last page. ** The document seems to lack a 1id_guidelines paragraph about Internet-Drafts being working documents. ** The document seems to lack a 1id_guidelines paragraph about 6 months document validity -- however, there's a paragraph with a matching beginning. Boilerplate error? ** The document seems to lack a 1id_guidelines paragraph about the list of current Internet-Drafts. ** The document seems to lack a 1id_guidelines paragraph about the list of Shadow Directories. == The page length should not exceed 58 lines per page, but there was 1 longer page, the longest (page 1) being 239 lines Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** The document seems to lack an Abstract section. ** The document seems to lack a Security Considerations section. ** The document seems to lack an IANA Considerations section. (See Section 2.2 of https://www.ietf.org/id-info/checklist for how to handle the case when there are no actions for IANA.) ** The document seems to lack an Authors' Addresses Section. ** The document seems to lack separate sections for Informative/Normative References. All references will be assumed normative when checking for downward references. ** There is 1 instance of too long lines in the document, the longest one being 2 characters in excess of 72. Miscellaneous warnings: ---------------------------------------------------------------------------- -- The document seems to lack a disclaimer for pre-RFC5378 work, but may have content which was first submitted before 10 November 2008. If you have contacted all the original authors and they are all willing to grant the BCP78 rights to the IETF Trust, then this is fine, and you can ignore this comment. If not, you may need to add the pre-RFC5378 disclaimer. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (Aug 1998) is 9383 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- -- Missing reference section? '1' on line 173 looks like a reference Summary: 12 errors (**), 0 flaws (~~), 1 warning (==), 3 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 INTERNET DRAFT EXPIRES FEB 1999 INTERNET DRAFT 4 End-To-End Research Group C. Partridge 5 INTERNET DRAFT BBN Technologies 6 Category: Informational Aug 1998 8 ACK Spacing for High Delay-Bandwidth Paths with Insufficient Buffering 9 11 Status of This Memo 13 This document is an Internet-Draft. Internet-Drafts are working 14 documents of the Internet Engineering Task Force (IETF), its areas, and 15 its working groups. Note that other groups may also distribute working 16 documents as Internet-Drafts. 18 Internet-Drafts are draft documents valid for a maximum of six months 19 and may be updated, replaced, or obsoleted by other documents at any 20 time. It is inappropriate to use Internet- Drafts as reference material 21 or to cite them other than as "work in progress." 23 To view the entire list of current Internet-Drafts, please check the 24 "1id-abstracts.txt" listing contained in the Internet-Drafts Shadow 25 Directories on ftp.is.co.za (Africa), ftp.nordu.net (Northern Europe), 26 ftp.nis.garr.it (Southern Europe), munnari.oz.au (Pacific Rim), 27 ftp.ietf.org (US East Coast), or ftp.isi.edu (US West Coast). 29 Distribution of this document is unlimited. 31 1. Introduction 33 Suppose you want TCP implementations to be able to fill a 155 Mb/s 34 path. Further suppose that the path includes a satellite in a 35 geosynchronous orbit, so the round trip delay through the path is at 36 least 500 ms, and the delay-bandwidth product is 9.7 megabytes or 37 more. 39 If we further assume the TCP implementations support TCP Large 40 Windows and PAWS (many do), so they can manage 9.7 MB TCP window, 41 then we can be sure the TCP will eventually start sending at full 42 path rate (unless the satellite channel is very lossy). But it may 43 take a long time to get the TCP up to full speed. 45 Partridge [Page 1]FORMFEED 46 One (of several) possible causes of the delay is a shortage of 47 buffering in routers. To understand this particular problem, 48 consider the following idealized behavior of TCP during slow start. 49 During slow start, for every segment ACKed, the sender transmits two 50 new segments. In effect, this behavior means the sender is 51 transmitting at *twice* the data rate of the segments being ACKed. 52 Keep in mind the separation between ACKs represents (in an ideal 53 world) the rate segments can flow through the bottleneck router in 54 the path. So the sender is bursting data at twice the bottleneck 55 rate, and a queue must be forming during the burst. In the simplest 56 case, the queue is entirely at the bottleneck router, and at the end 57 of the burst, the queue is storing half the data in the burst. (Why 58 half? During the burst, the sender transmitted at twice the 59 bottleneck rate. Suppose it takes one time unit to send a segment on 60 the bottlenecked link. During the burst the bottleneck will receive 61 two segments in every time unit, but only be able to transmit one 62 segment. The result is a net of one new segment queued every time 63 unit, for the life of the burst.) 65 TCP will end the slow start phase in response to the first lost 66 datagram. Assuming good quality transmission links, the first lost 67 datagram will be lost because the bottleneck queue overflowed. We 68 would like that loss to occur in the round-trip after the slow start 69 congestion window has reached the delay-bandwidth product. Now 70 consider the buffering required in the bottleneck link during the 71 next to last round trip. The sender will send an entire delay- 72 bandwidth worth of data in one-half a round-trip time (because it 73 sends at twice the channel rate). So for half the round-trip time, 74 the bottleneck router is in the mode of forwarding one segment while 75 receiving two. (For the second half of the round-trip, the router is 76 draining its queue). That means, to avoid losing any segments, the 77 router must have buffering equal to half the delay-bandwidth product, 78 or nearly 5 MB. 80 Most routers do not have anywhere near 5 MB of buffering for a single 81 link. Or, to express this problem another way, because routers do 82 not have this much buffering, the slow start stage will end 83 prematurely, when router buffering is exhausted. The consequence of 84 ending slow start prematurely is severe. At the end of slow start, 85 TCP goes into congestion avoidance, in which the window size is 86 increased much more slowly. So even though the channel is free, 87 because we did not have enough router buffering, we will transmit 88 slowly for a period of time (until the more conservative congestion 89 avoidance algorithm sends enough data to fill the channel). 91 Partridge [Page 2]FORMFEED 92 2. What to Do? 94 So how to get around the shortage of router buffering? 96 One solution has been proposed, cascading TCPs. We would like to 97 suggest another solution, ACK spacing. Both schemes involve layer 98 violations because they require the router to examine the TCP header. 100 2.1 Cascading TCPs 102 One approach is to use cascading TCPs, in which we build a custom TCP 103 for the satellite (or bottleneck) link and insert it between the 104 sender's and receiver's TCPs, as shown below: 106 sender ---- Ground station -- satellite -- ground station -- receiver 108 +---------------+ +------------------------+ +---------------+ 109 | loop 1 | | loop 2 | | loop 3 | 110 +---------------+ +------------------------+ +---------------+ 112 This approach can work but is awkward. Among its limitations are: 113 the buffering problem remains (at points of bandwidth mismatches, 114 queues will form); the scheme violates end-to-end semantics of TCP 115 (the sender will get ACKs for data that has not and may never reach 116 the receiver); it constrains the reverse path of the TCP connection 117 to pass through points at which the multiple TCP connections are 118 spliced together (a problem if satellite links are unidirectional); 119 and it doesn't work with end-to-end encryption (i.e. if data above 120 the IP layer is encrypted). 122 2.2 ACK Spacing 124 Another approach is to find some way to spread the bursts, either by 125 having the sender spread out the segments, or having the network 126 arrange for the ACKs to arrive at the sender with a two segment 127 spacing (or larger). 129 Changing the sender is feasible, although it requires very good 130 operating system timers. But it has the disadvantage that only 131 upgraded senders get the performance improvement. 133 Finding a way for the network to space the ACKs would allow TCP 134 senders to transmit at the right rate, without modification. 135 Furthermore, it can be done by a router. The router simply has to 136 snoop the returning TCP ACKs and spread them out. (Note that if the 137 transmissions are encrypted, in many scenarios the router can still 138 figure out which segments are likely TCP ACKs and spread them out). 140 Partridge [Page 3]FORMFEED 141 There are some difficult issues with this approach. The most notable 142 ones are: 144 1. What algorithm to use to determine the proper ACK spacing. 146 2. Related to (1), it may be necessary to known when a TCP is in 147 slow-start vs. congestion-avoidance, as the desired spacing 148 between ACKs is likely to be different in the two phases. 150 3. What to do about assymetric routes (if anything). The scheme 151 works so long as the router sees the ACKs (it does not have to see 152 the related data). However, if the ACKs do not return through the 153 ACK-spacing router, it is not possible to do ACK spacing. 155 4. How much, if at all, does ack compression between the respacing 156 point and the sender undo the effects of ack spacing? 158 5. How much per-flow (soft) state is required in the ACK spacing 159 router? 161 Despite these challenges the approach has appeal. Changing software 162 in a few routers (particularly those at likely bottleneck links) on 163 high delay-bandwidth paths could give a performance boost to lots of 164 TCP connections. 166 Credit and Disclaimer 168 This memo presents thoughts from a discussion held at the recent 169 meeting of the End-To-End (E2E) Research Group. The particular idea 170 of ACK spacing was developed by during the meeting by Mark Handley 171 and Van Jacobson in response to an issue raised by the author, and 172 was inspired, in part by ideas to enhance wireless routers to improve 173 TCP performance [1]. 175 The material presented is a half-baked suggestion and should not be 176 interpreted as an official recommendation of the Research Group. 178 References 180 1. H. Balakrishnan, V.N. Padmanabhan, S. Seshan and R.H. Katz, "A 181 Comparison of Mechanisms for Improving TCP Performance over Wireless 182 Links", Proc. ACM SIGCOMM '96, pp. 256-269. 184 DRAFT 186 Partridge [Page 4]FORMFEED