1658:...The advantage of this function lies not only in avoiding heavy oscillations but also in avoiding link under-utilization at low loads. The applicability of the derived function is independent of the load range, no parameters are to be adjusted. Compared to the original linear drop function applicability is extended by far...Our example with realistic system parameters gives an approximation function of the cubic of the queue size...
261:
73:
to compensate for packet loss due to congestion can increase congestion, even after the initial load has been reduced to a level that would not normally have induced network congestion. Such networks exhibit two stable states under the same level of load. The stable state with low throughput is known
1391:
In
October of '86, the Internet had the first of what became a series of 'congestion collapses'. During this period, the data throughput from LBL to UC Berkeley (sites separatedby 400 yards and two IMP hops) dropped from 32 Kbps to 40 bps. We were fascinated by this sudden factor-of-thousand drop in
1131:
When a router receives a packet marked as ECN-capable and the router anticipates congestion, it sets the ECN flag, notifying the sender of congestion. The sender should respond by decreasing its transmission bandwidth, e.g., by decreasing its sending rate by reducing the TCP window size or by other
1107:(RRED) algorithm was proposed to improve the TCP throughput against denial-of-service (DoS) attacks, particularly low-rate denial-of-service (LDoS) attacks. Experiments confirmed that RED-like algorithms were vulnerable under LDoS attacks due to the oscillating TCP queue size caused by the attacks.
1163:
messages as an IP signaling mechanism to implement a basic ECN mechanism for IP networks, keeping congestion notifications at the IP level and requiring no negotiation between network endpoints. Effective congestion notifications can be propagated to transport layer protocols, such as TCP and UDP,
1150:
of between 32K and 64K. This results in the server sending a full window of data (assuming the file is larger than the window). When many applications simultaneously request downloads, this data can create a congestion point at an upstream provider. By reducing the window advertisement, the remote
1201:
created many short-lived connections and opened and closed the connection for each file. This kept most connections in the slow start mode. Initial performance can be poor, and many connections never get out of the slow-start regime, significantly increasing latency. To avoid this problem, modern
225:
were sent than could be handled by intermediate routers, the intermediate routers discarded many packets, expecting the end points of the network to retransmit the information. However, early TCP implementations had poor retransmission behavior. When this packet loss occurred, the endpoints sent
1128:(ECN). ECN is used only when two hosts signal that they want to use it. With this method, a protocol bit is used to signal explicit congestion. This is better than the indirect congestion notification signaled by packet loss by the RED/WRED algorithms, but it requires support by both hosts.
234:
Congestion control modulates traffic entry into a telecommunications network in order to avoid congestive collapse resulting from oversubscription. This is typically accomplished by reducing the rate of packets. Whereas congestion control prevents senders from overwhelming the
1392:
bandwidth and embarked on an investigation of why things had gotten so bad. In particular, we wondered if the 4.3BSD(Berkeley UNIX)TCPwas mis-behaving or if it could be tuned to work better under abysmal network conditions.The answer to both of these questions was "yes".
189:
Congestive collapse (or congestion collapse) is the condition in which congestion prevents or limits useful communication. Congestion collapse generally occurs at choke points in the network, where incoming traffic exceeds outgoing bandwidth. Connection points between a
1217:
is any system that requires devices to receive permission before establishing new network connections. If the new connection risks creating congestion, permission can be denied. Examples include
Contention-Free Transmission Opportunities (CFTXOPs) in the ITU-T
1115:
Some network equipment is equipped with ports that can follow and measure each flow and are thereby able to signal a too big bandwidth flow according to some quality of service policy. A policy could then divide the bandwidth among all flows by some criteria.
953:
The correct endpoint behavior is usually to repeat dropped information, but progressively slow the repetition rate. Provided all endpoints do this, the congestion lifts and the network resumes normal behavior. Other strategies such as
990:
does not control congestion. Protocols built atop UDP must handle congestion independently. Protocols that transmit at a fixed rate, independent of congestion, can be problematic. Real-time streaming protocols, including many
129:. Other techniques that address congestion include priority schemes which transmit some packets with higher priority ahead of others and the explicit allocation of network resources to specific flows through the use of
1312:
980:
is a well known example. The first TCP implementations to handle congestion were described in 1984, but Van
Jacobson's inclusion of an open source solution in the Berkeley Standard Distribution UNIX ("
690:
1041:
is present. This delayed packet loss interferes with TCP's automatic congestion avoidance. All flows that experience this packet loss begin a TCP retrain at the same moment – this is called
831:
1185:
and other networks with a radio layer are susceptible to data loss due to interference and may experience poor throughput in some cases. The TCP connections running over a radio-based
908:
By incremental deployability: Only sender needs modification; sender and receiver need modification; only router needs modification; sender, receiver and routers need modification.
198:
are common choke points. When a network is in this condition, it settles into a stable state where traffic demand is high but little useful throughput is available, during which
1943:
Pierre-Francois Quet, Sriram
Chellappan, Arjan Durresi, Mukundan Sridharan, Hitay Ozbay, Raj Jain, "Guidelines for optimizing Multi-Level ECN, using fluid flow based TCP model"
1671:
723:
474:
973:(RED) where packets are randomly dropped as congestion is detected. This proactively triggers the endpoints to slow transmission before congestion collapse occurs.
861:
765:
424:
377:
995:
protocols, have this property. Thus, special measures, such as quality of service, must be taken to keep packets from being dropped in the presence of congestion.
603:
881:
631:
574:
554:
534:
514:
494:
444:
397:
1177:
The protocols that avoid congestive collapse generally assume that data loss is caused by congestion. On wired networks, errors during transmission are rare.
1146:
Congestion avoidance can be achieved efficiently by reducing traffic. When an application requests a large file, graphic or web page, it usually advertises a
213:
phase-I backbone dropped three orders of magnitude from its capacity of 32 kbit/s to 40 bit/s, which continued until end nodes started implementing
1087:
RED indirectly signals TCP sender and receiver by dropping some packets, e.g. when the average queue length is more than a threshold (e.g. 50%) and deletes
1470:
836:
Congestion control then becomes a distributed optimization algorithm. Many current congestion control algorithms can be modelled in this framework, with
2029:
1716:
1969:
1325:
1838:
1803:
1564:, vol.18(4): pp.314–329. Stanford, CA. August, 1988. This paper originated many of the congestion avoidance algorithms used in TCP/IP.
282:
209:
Congestive collapse was identified as a possible problem by 1984. It was first observed on the early
Internet in October 1986, when the
2019:
1989:
1027:
1527:
1364:
1291:
1003:
308:
1267:
911:
By performance aspect: high bandwidth-delay product networks; lossy links; fairness; advantage to short flows; variable-rate links
1784:
Pop, O.; Moldován, I.; Simon, Cs.; Bíró, J.; Koike, A.; Ishii, H. (2000), "Advertised Window-Based TCP Flow
Control in Routers",
1125:
940:
2024:
1081:
222:
1950:
1994:
286:
70:
1438:. 16th IFAC Workshop on Distributed Computer Control Systems (DCCS 2000), Sydney, Australia, 29 November-1 December 2000.
1223:
1104:
1007:
977:
114:
1964:
2009:
1279:
1057:
639:
1933:
1227:
1203:
271:
1600:
1510:
Lee, B.P.; Balan, R.K.; Jacob, L.; Seah, W.K.G.; Ananda, A.L. (2000), "TCP Tunnels: Avoiding
Congestion Collapse",
1042:
290:
275:
770:
1324:
den Hartog, F., Raschella, A., Bouhafs, F., Kempker, P., Boltjes, B., & Seyedebrahimi, M. (2017, November).
1018:
to adjust their transmission rate. Various network congestion avoidance processes support different trade-offs.
905:
By type and amount of feedback received from the network: Loss; delay; single-bit or multi-bit explicit signals
178:
162:
1493:
1328:. In 2017 27th International Telecommunication Networks and Applications Conference (ITNAC) (pp. 1-6). IEEE.
1273:
1141:
1053:
934:
883:. A major weakness is that it assigns the same price to all flows, while sliding window flow control causes
322:
1760:
1647:
1073:
1056:(AQM) is the reordering or dropping of network packets inside a transmit buffer that is associated with a
987:
970:
955:
946:
896:
47:
that occurs when a network node or link is carrying more data than it can handle. Typical effects include
943: – an extension to IP and TCP communications protocols that adds a flow control mechanism
966:
1431:
915:
2014:
1916:
1243:
326:
1652:
1405:
1077:
737:
346:
330:
240:
150:
90:
55:
or the blocking of new connections. A consequence of congestion is that an incremental increase in
1698:
1533:
1160:
884:
203:
191:
142:
122:
44:
926:
Mechanisms have been invented to prevent network congestion or to deal with a network collapse:
1834:
1799:
1573:
RFC 2001 - TCP Slow Start, Congestion
Avoidance, Fast Retransmit, and Fast Recovery Algorithms
1553:
1523:
1451:
1360:
1249:
1214:
1061:
958:
ensure that new connections don't overwhelm the router before congestion detection initiates.
930:
699:
195:
170:
130:
1872:
1853:
1789:
1690:
1515:
1485:
1443:
1337:
729:
606:
449:
342:
158:
66:
839:
743:
402:
355:
1954:
1937:
1285:
579:
333:
theory to describe how individuals controlling their own rates can interact to achieve an
174:
40:
36:
1917:
On the
Evolution of End-to-end Congestion Control in the Internet: An Idiosyncratic View
1817:
736:
signaled by the network. Each link capacity imposes a constraint, which gives rise to a
1255:
1186:
1092:
866:
616:
559:
539:
519:
499:
479:
429:
382:
126:
110:
48:
1942:
1748:
1447:
2003:
1720:
1261:
1015:
992:
949: – various implementations of efforts to deal with network congestion
118:
1537:
157:
is easily filled by a single personal computer. Even on fast computer networks, the
1948:
Sally Floyd, Ratul
Mahajan, David Wetherall: RED-PD: RED with Preferential Dropping
1947:
1702:
1549:
1159:
Backward ECN (BECN) is another proposed congestion notification mechanism. It uses
962:
214:
199:
173:
network links, generating large-scale network congestion. In telephone networks, a
154:
56:
17:
976:
Some end-to-end protocols are designed to behave well under congested conditions;
1794:
1694:
1354:
937:
which reorders or selectively drops network packets in the presence of congestion
1907:
1198:
1189:
see the data loss and tend to erroneously believe that congestion is occurring.
1038:
1011:
887:
that causes different flows to observe different loss or delay at a given link.
260:
218:
177:
can overwhelm digital telephone circuits, in what can otherwise be defined as a
52:
1341:
732:
of this problem decouples so that each flow sets its own rate, based only on a
1489:
226:
extra packets that repeated the information lost, doubling the incoming rate.
146:
60:
1920:(IMA Workshop on Scaling Phenomena in Communication Networks, October 1999) (
1633:, vol.1(4): pp.397–413. Invented Random Early Detection (RED) gateways.
1455:
1258: – Telephone exchange designed to handle many simultaneous call attempts
1930:
1519:
1512:
Proceedings 25th Annual IEEE Conference on Local Computer Networks. LCN 2000
1034:
1672:"RRED: Robust RED Algorithm to Counter Low-rate Denial-of-Service Attacks"
1626:
1557:
1381:
1887:
1326:
A Pathway to solving the Wi-Fi Tragedy of the Commons in apartment blocks
1231:
613:, which measures how much benefit a user obtains by transmitting at rate
106:
1831:
Deploying IP and MPLS QoS for Multiservice Networks: Theory and Practice
1960:
A Generic Simple RED Simulator for educational purposes by Mehmet Suzen
1739:
RFC 3168 - The Addition of Explicit Congestion Notification (ECN) to IP
1276: – Application of traffic engineering theory to telecommunications
610:
102:
94:
1858:
1979:
1974:
1877:
1670:
Zhang, Changwang; Yin, Jianping; Cai, Zhiping; Chen, Weifeng (2010).
1644:
An Analytical RED Function Design Guaranteeing Stable System Behavior
1614:
1088:
210:
166:
98:
1313:
Simulation in computer network design and modeling: Use and analysis
1197:
The slow-start protocol performs badly for short connections. Older
59:
leads either only to a small increase or even a decrease in network
1959:
1908:
Promoting the Use of End-to-End Congestion Control in the Internet
1432:"A Control Theory Approach for Congestion Control in Intranetwork"
1686:
1406:"Sally Floyd, Who Helped Things Run Smoothly Online, Dies at 69"
1219:
1178:
863:
being either the loss probability or the queueing delay at link
1995:
Recent Publications in low-rate denial-of-service (DoS) attacks
1922:
1270: – Network protocol flaw in the original versions of TFTP
981:
914:
By fairness criterion: Max-min fairness; proportionally fair;
901:
Among the ways to classify congestion control algorithms are:
254:
1030:
is the primary basis for congestion control on the Internet.
1984:
1202:
browsers either open multiple connections simultaneously or
1182:
1095:
more packets, up to e.g. 100%, as the queue fills further.
1615:
Sally Floyd: RED (Random Early Detection) Queue Management
1749:
Comparative study of RED, ECN and TCP Rate Control (1999)
1601:
TCP Congestion Avoidance Explained via a Sequence Diagram
161:
can easily be congested by a few servers and client PCs.
153:
may occur on networks in several common circumstances. A
1762:
Generalized Window Advertising for TCP CongestionControl
1627:
Random Early Detection Gateways for Congestion Avoidance
221:'s congestion control between 1987 and 1988. When more
1151:
servers send less data, thus reducing the congestion.
961:
Common router congestion avoidance mechanisms include
27:
Reduced quality of service due to high network traffic
1818:
A proposal for Backward ECN for the Internet Protocol
1311:(Al-Bahadili, 2012, p. 282) Al-Bahadili, H. (2012).
869:
842:
773:
746:
702:
642:
619:
582:
562:
542:
522:
502:
482:
452:
432:
405:
385:
358:
1288: – Communication bandwidth management technique
1282: – Constant exchange between memory and storage
1246: – Capacity control on a communications network
1033:
Problems occur when concurrent TCP flows experience
89:
techniques to try to avoid collapse. These include:
1965:
Approaches to Congestion Control in Packet Networks
1911:(IEEE/ACM Transactions on Networking, August 1999)
1471:"A Protocol for Packet Network Intercommunication"
1206:for all files requested from a particular server.
1076:(RED) on the network equipment's egress queue. On
875:
855:
825:
759:
717:
685:{\displaystyle \max \limits _{x}\sum _{i}U(x_{i})}
684:
625:
597:
568:
548:
528:
508:
488:
468:
438:
418:
391:
371:
321:The theory of congestion control was pioneered by
1886:Van Jacobson; Michael J. Karels (November 1988).
1380:Van Jacobson; Michael J. Karels (November 1988),
1222:standard for home networking over legacy wiring,
891:Classification of congestion control algorithms
349:allocation, although many others are possible.
1591:RFC 3390 - TCP Increasing TCP's Initial Window
1359:(2 ed.). Pearson Education. p. 739.
1168:Side effects of congestive collapse avoidance
633:. The optimal rate allocation then satisfies
576:be the corresponding vectors and matrix. Let
8:
1610:
1608:
1469:Vinton G. Cerf; Robert E. Kahn (May 1974).
1356:TCP/IP Illustrated, Volume 1: The Protocols
1294: – Protocol acknowledgement capability
826:{\displaystyle y_{i}=\sum _{l}p_{l}r_{li},}
289:. Unsourced material may be challenged and
1264: – Load measure in telecommunications
337:network-wide rate allocation. Examples of
243:prevents the sender from overwhelming the
1980:Explicit Congestion Notification Homepage
1876:
1857:
1793:
1651:
984:") in 1988 first provided good behavior.
868:
847:
841:
833:is the price to which the flow responds.
811:
801:
791:
778:
772:
751:
745:
701:
673:
657:
647:
641:
618:
581:
561:
541:
521:
501:
481:
457:
451:
431:
410:
404:
384:
363:
357:
309:Learn how and when to remove this message
141:Network resources are limited, including
1562:Proceedings of the Sigcomm '88 Symposium
169:are capable of filling even the largest
1304:
1080:ports with more than one egress queue,
1829:John Evans; Clarence Filsfils (2007).
1786:Telecommunication Network Intelligence
999:Practical network congestion avoidance
1060:(NIC). This task is performed by the
7:
287:adding citations to reliable sources
1631:IEEE/ACM Transactions on Networking
1478:IEEE Transactions on Communications
1888:"Congestion Avoidance and Control"
1404:Hafner, Katie (4 September 2019).
1353:Fall, K.R.; Stevens, W.R. (2011).
1028:TCP congestion avoidance algorithm
25:
1582:RFC 2581 - TCP Congestion Control
1292:Reliability (computer networking)
1164:for the appropriate adjustments.
2030:Packets (information technology)
1558:Congestion Avoidance and Control
1430:Nanda, Priyadarsi (2000-11-01).
1383:Congestion Avoidance and Control
1252: – Systemic risk of failure
1126:Explicit Congestion Notification
1120:Explicit Congestion Notification
941:Explicit Congestion Notification
767:. The sum of these multipliers,
259:
1975:Random Early Detection Homepage
1717:"Congestion Avoidance Overview"
1082:weighted random early detection
1848:Sally Floyd (September 2000).
1268:Sorcerer's Apprentice syndrome
679:
666:
592:
586:
1:
1867:John Nagle (6 January 1984).
1850:Congestion Control Principles
1448:10.1016/S1474-6670(17)36735-6
1224:Resource Reservation Protocol
1105:robust random early detection
1099:Robust random early detection
1004:Connection-oriented protocols
1970:Papers in Congestion Control
1869:Congestion Control in IP/TCP
1795:10.1007/978-0-387-35522-1_12
1695:10.1109/LCOMM.2010.05.091407
1058:network interface controller
251:Theory of congestion control
1679:IEEE Communications Letters
1625:Sally Floyd, Van Jacobson.
1315:. Hershey, PA: IGI Global.
1228:Stream Reservation Protocol
1124:Another approach is to use
1022:TCP/IP congestion avoidance
644:
605:be an increasing, strictly
2046:
1139:
1043:TCP global synchronization
1006:, such as the widely used
894:
345:and Kelly's suggestion of
2020:Transport layer protocols
1931:Linktionary term: Queuing
1490:10.1109/tcom.1974.1092259
163:Denial-of-service attacks
145:processing time and link
1436:IFAC Proceedings Volumes
718:{\displaystyle Rx\leq c}
426:be the capacity of link
179:denial-of-service attack
1905:Floyd, S. and K. Fall,
1520:10.1109/LCN.2000.891077
1274:Teletraffic engineering
1193:Short-lived connections
1142:TCP window scale option
1072:One solution is to use
1054:Active queue management
1049:Active queue management
935:active queue management
343:max-min fair allocation
2025:Technological failures
1074:random early detection
1068:Random early detection
971:random early detection
947:TCP congestion control
897:TCP congestion control
877:
857:
827:
761:
719:
686:
627:
599:
570:
550:
530:
510:
490:
470:
469:{\displaystyle r_{li}}
440:
420:
393:
373:
967:scheduling algorithms
878:
858:
856:{\displaystyle p_{l}}
828:
762:
760:{\displaystyle p_{l}}
720:
687:
628:
600:
571:
551:
531:
516:and 0 otherwise. Let
511:
491:
471:
441:
421:
419:{\displaystyle c_{l}}
394:
374:
372:{\displaystyle x_{i}}
93:in protocols such as
1788:, pp. 197–218,
1514:, pp. 408–417,
1244:Bandwidth management
1226:for IP networks and
1204:reuse one connection
1084:(WRED) can be used.
1010:protocol, watch for
867:
840:
771:
744:
700:
640:
617:
598:{\displaystyle U(x)}
580:
560:
540:
520:
500:
480:
450:
430:
403:
383:
379:be the rate of flow
356:
341:rate allocation are
327:microeconomic theory
283:improve this section
87:congestion avoidance
69:that use aggressive
2010:Network performance
1833:. Morgan Kaufmann.
1078:networking hardware
933: –
738:Lagrange multiplier
347:proportionally fair
331:convex optimization
206:is extremely poor.
202:and loss occur and
185:Congestive collapse
151:Resource contention
121:in devices such as
91:exponential backoff
76:congestive collapse
18:Congestion collapse
1953:2003-04-02 at the
1936:2003-03-08 at the
1161:ICMP source quench
1136:TCP window shaping
1037:, especially when
873:
853:
823:
796:
757:
715:
682:
662:
652:
623:
595:
566:
546:
526:
506:
486:
466:
436:
416:
389:
369:
230:Congestion control
204:quality of service
192:local area network
83:congestion control
45:quality of service
33:Network congestion
1840:978-0-12-370549-5
1805:978-1-4757-6693-6
1554:Michael J. Karels
1499:on March 4, 2016.
1250:Cascading failure
1215:Admission control
1210:Admission control
1062:network scheduler
931:Network scheduler
876:{\displaystyle l}
787:
653:
643:
626:{\displaystyle x}
569:{\displaystyle R}
549:{\displaystyle c}
529:{\displaystyle x}
509:{\displaystyle l}
489:{\displaystyle i}
439:{\displaystyle l}
392:{\displaystyle i}
319:
318:
311:
196:wide area network
171:Internet backbone
131:admission control
67:Network protocols
16:(Redirected from
2037:
1990:AIMD-FC Homepage
1894:
1892:
1882:
1880:
1863:
1861:
1844:
1820:
1815:
1809:
1808:
1797:
1781:
1775:
1774:
1773:
1772:
1767:
1757:
1751:
1746:
1740:
1737:
1731:
1730:
1728:
1727:
1713:
1707:
1706:
1676:
1667:
1661:
1660:
1655:
1640:
1634:
1623:
1617:
1612:
1603:
1598:
1592:
1589:
1583:
1580:
1574:
1571:
1565:
1547:
1541:
1540:
1507:
1501:
1500:
1498:
1492:. Archived from
1475:
1466:
1460:
1459:
1427:
1421:
1420:
1418:
1416:
1401:
1395:
1394:
1388:
1377:
1371:
1370:
1350:
1344:
1335:
1329:
1322:
1316:
1309:
916:controlled delay
882:
880:
879:
874:
862:
860:
859:
854:
852:
851:
832:
830:
829:
824:
819:
818:
806:
805:
795:
783:
782:
766:
764:
763:
758:
756:
755:
724:
722:
721:
716:
691:
689:
688:
683:
678:
677:
661:
651:
632:
630:
629:
624:
607:concave function
604:
602:
601:
596:
575:
573:
572:
567:
555:
553:
552:
547:
535:
533:
532:
527:
515:
513:
512:
507:
495:
493:
492:
487:
475:
473:
472:
467:
465:
464:
445:
443:
442:
437:
425:
423:
422:
417:
415:
414:
398:
396:
395:
390:
378:
376:
375:
370:
368:
367:
314:
307:
303:
300:
294:
263:
255:
137:Network capacity
127:network switches
105:in the original
101:and the similar
21:
2045:
2044:
2040:
2039:
2038:
2036:
2035:
2034:
2000:
1999:
1955:Wayback Machine
1938:Wayback Machine
1902:
1897:
1890:
1885:
1866:
1847:
1841:
1828:
1824:
1823:
1816:
1812:
1806:
1783:
1782:
1778:
1770:
1768:
1765:
1759:
1758:
1754:
1747:
1743:
1738:
1734:
1725:
1723:
1715:
1714:
1710:
1674:
1669:
1668:
1664:
1653:10.1.1.105.5995
1642:
1641:
1637:
1624:
1620:
1613:
1606:
1599:
1595:
1590:
1586:
1581:
1577:
1572:
1568:
1548:
1544:
1530:
1509:
1508:
1504:
1496:
1473:
1468:
1467:
1463:
1429:
1428:
1424:
1414:
1412:
1403:
1402:
1398:
1386:
1379:
1378:
1374:
1367:
1352:
1351:
1347:
1336:
1332:
1323:
1319:
1310:
1306:
1301:
1286:Traffic shaping
1240:
1212:
1195:
1175:
1170:
1157:
1144:
1138:
1122:
1113:
1111:Flow-based WRED
1101:
1070:
1051:
1024:
1001:
924:
899:
893:
865:
864:
843:
838:
837:
807:
797:
774:
769:
768:
747:
742:
741:
698:
697:
669:
638:
637:
615:
614:
578:
577:
558:
557:
538:
537:
518:
517:
498:
497:
478:
477:
453:
448:
447:
428:
427:
406:
401:
400:
381:
380:
359:
354:
353:
315:
304:
298:
295:
280:
264:
253:
232:
187:
175:mass call event
139:
71:retransmissions
43:is the reduced
41:queueing theory
37:data networking
28:
23:
22:
15:
12:
11:
5:
2043:
2041:
2033:
2032:
2027:
2022:
2017:
2012:
2002:
2001:
1998:
1997:
1992:
1987:
1982:
1977:
1972:
1967:
1962:
1957:
1945:
1940:
1928:
1912:
1901:
1900:External links
1898:
1896:
1895:
1883:
1864:
1845:
1839:
1825:
1822:
1821:
1810:
1804:
1776:
1752:
1741:
1732:
1708:
1662:
1635:
1618:
1604:
1593:
1584:
1575:
1566:
1542:
1528:
1502:
1484:(5): 637–648.
1461:
1422:
1410:New York Times
1396:
1372:
1365:
1345:
1330:
1317:
1303:
1302:
1300:
1297:
1296:
1295:
1289:
1283:
1277:
1271:
1265:
1259:
1256:Choke exchange
1253:
1247:
1239:
1236:
1211:
1208:
1194:
1191:
1187:physical layer
1174:
1171:
1169:
1166:
1156:
1153:
1137:
1134:
1121:
1118:
1112:
1109:
1100:
1097:
1069:
1066:
1050:
1047:
1023:
1020:
1000:
997:
951:
950:
944:
938:
923:
920:
919:
918:
912:
909:
906:
892:
889:
872:
850:
846:
822:
817:
814:
810:
804:
800:
794:
790:
786:
781:
777:
754:
750:
726:
725:
714:
711:
708:
705:
693:
692:
681:
676:
672:
668:
665:
660:
656:
650:
646:
622:
594:
591:
588:
585:
565:
545:
525:
505:
485:
463:
460:
456:
435:
413:
409:
388:
366:
362:
325:, who applied
317:
316:
267:
265:
258:
252:
249:
231:
228:
186:
183:
138:
135:
49:queueing delay
26:
24:
14:
13:
10:
9:
6:
4:
3:
2:
2042:
2031:
2028:
2026:
2023:
2021:
2018:
2016:
2013:
2011:
2008:
2007:
2005:
1996:
1993:
1991:
1988:
1986:
1985:TFRC Homepage
1983:
1981:
1978:
1976:
1973:
1971:
1968:
1966:
1963:
1961:
1958:
1956:
1952:
1949:
1946:
1944:
1941:
1939:
1935:
1932:
1929:
1926:
1924:
1919:
1918:
1914:Sally Floyd,
1913:
1910:
1909:
1904:
1903:
1899:
1889:
1884:
1879:
1874:
1870:
1865:
1860:
1855:
1851:
1846:
1842:
1836:
1832:
1827:
1826:
1819:
1814:
1811:
1807:
1801:
1796:
1791:
1787:
1780:
1777:
1764:
1763:
1756:
1753:
1750:
1745:
1742:
1736:
1733:
1722:
1721:Cisco Systems
1718:
1712:
1709:
1704:
1700:
1696:
1692:
1688:
1684:
1680:
1673:
1666:
1663:
1659:
1654:
1649:
1645:
1639:
1636:
1632:
1628:
1622:
1619:
1616:
1611:
1609:
1605:
1602:
1597:
1594:
1588:
1585:
1579:
1576:
1570:
1567:
1563:
1559:
1555:
1551:
1546:
1543:
1539:
1535:
1531:
1529:0-7695-0912-6
1525:
1521:
1517:
1513:
1506:
1503:
1495:
1491:
1487:
1483:
1479:
1472:
1465:
1462:
1457:
1453:
1449:
1445:
1442:(30): 91–94.
1441:
1437:
1433:
1426:
1423:
1411:
1407:
1400:
1397:
1393:
1385:
1384:
1376:
1373:
1368:
1366:9780132808187
1362:
1358:
1357:
1349:
1346:
1343:
1339:
1334:
1331:
1327:
1321:
1318:
1314:
1308:
1305:
1298:
1293:
1290:
1287:
1284:
1281:
1278:
1275:
1272:
1269:
1266:
1263:
1262:Erlang (unit)
1260:
1257:
1254:
1251:
1248:
1245:
1242:
1241:
1237:
1235:
1233:
1229:
1225:
1221:
1216:
1209:
1207:
1205:
1200:
1192:
1190:
1188:
1184:
1180:
1172:
1167:
1165:
1162:
1154:
1152:
1149:
1143:
1135:
1133:
1129:
1127:
1119:
1117:
1110:
1108:
1106:
1098:
1096:
1094:
1090:
1085:
1083:
1079:
1075:
1067:
1065:
1063:
1059:
1055:
1048:
1046:
1044:
1040:
1036:
1031:
1029:
1021:
1019:
1017:
1016:queuing delay
1013:
1009:
1005:
998:
996:
994:
993:Voice over IP
989:
985:
983:
979:
974:
972:
968:
964:
959:
957:
948:
945:
942:
939:
936:
932:
929:
928:
927:
921:
917:
913:
910:
907:
904:
903:
902:
898:
890:
888:
886:
870:
848:
844:
834:
820:
815:
812:
808:
802:
798:
792:
788:
784:
779:
775:
752:
748:
739:
735:
731:
730:Lagrange dual
712:
709:
706:
703:
695:
694:
674:
670:
663:
658:
654:
648:
636:
635:
634:
620:
612:
609:, called the
608:
589:
583:
563:
543:
523:
503:
483:
476:be 1 if flow
461:
458:
454:
433:
411:
407:
386:
364:
360:
350:
348:
344:
340:
336:
332:
328:
324:
313:
310:
302:
292:
288:
284:
278:
277:
273:
268:This section
266:
262:
257:
256:
250:
248:
246:
242:
238:
229:
227:
224:
220:
216:
212:
207:
205:
201:
197:
193:
184:
182:
180:
176:
172:
168:
164:
160:
156:
152:
148:
144:
136:
134:
132:
128:
124:
120:
119:fair queueing
116:
113:reduction in
112:
108:
104:
100:
96:
92:
88:
84:
81:Networks use
79:
77:
72:
68:
64:
62:
58:
54:
50:
46:
42:
38:
34:
30:
19:
1921:
1915:
1906:
1868:
1849:
1830:
1813:
1785:
1779:
1769:, retrieved
1761:
1755:
1744:
1735:
1724:. Retrieved
1711:
1682:
1678:
1665:
1657:
1643:
1638:
1630:
1621:
1596:
1587:
1578:
1569:
1561:
1550:Van Jacobson
1545:
1511:
1505:
1494:the original
1481:
1477:
1464:
1439:
1435:
1425:
1413:. Retrieved
1409:
1399:
1390:
1382:
1375:
1355:
1348:
1333:
1320:
1307:
1213:
1199:web browsers
1196:
1176:
1158:
1155:Backward ECN
1147:
1145:
1130:
1123:
1114:
1102:
1086:
1071:
1052:
1032:
1025:
1002:
986:
975:
963:fair queuing
960:
952:
925:
900:
835:
733:
727:
351:
338:
334:
320:
305:
296:
281:Please help
269:
244:
241:flow control
236:
233:
215:Van Jacobson
208:
200:packet delay
188:
155:wireless LAN
140:
86:
82:
80:
75:
65:
57:offered load
32:
31:
29:
2015:Teletraffic
1689:: 489–491.
1415:5 September
1173:Radio links
1039:bufferbloat
1012:packet loss
323:Frank Kelly
219:Sally Floyd
53:packet loss
2004:Categories
1771:2020-11-13
1726:2020-08-07
1299:References
1140:See also:
1035:tail-drops
965:and other
956:slow start
922:Mitigation
895:See also:
885:burstiness
696:such that
496:uses link
147:throughput
61:throughput
1648:CiteSeerX
1456:1474-6670
1280:Thrashing
1093:cubically
789:∑
710:≤
655:∑
270:does not
1951:Archived
1934:Archived
1629:(1993).
1560:(1988).
1538:34447400
1238:See also
1232:Ethernet
1089:linearly
299:May 2013
245:receiver
159:backbone
107:Ethernet
1703:1121461
1132:means.
611:utility
339:optimal
335:optimal
291:removed
276:sources
237:network
223:packets
167:botnets
123:routers
103:CSMA/CD
95:CSMA/CA
1925:format
1875:
1856:
1837:
1802:
1701:
1650:
1536:
1526:
1454:
1363:
1340:
1148:window
969:, and
446:, and
211:NSFNET
194:and a
143:router
117:, and
111:window
99:802.11
1891:(PDF)
1766:(PDF)
1699:S2CID
1685:(5).
1675:(PDF)
1534:S2CID
1497:(PDF)
1474:(PDF)
1387:(PDF)
734:price
1859:2914
1835:ISBN
1800:ISBN
1687:IEEE
1524:ISBN
1452:ISSN
1417:2019
1361:ISBN
1230:for
1220:G.hn
1179:WiFi
1103:The
1026:The
728:The
556:and
352:Let
329:and
274:any
272:cite
217:and
125:and
85:and
39:and
1923:pdf
1878:896
1873:RFC
1854:RFC
1790:doi
1691:doi
1516:doi
1486:doi
1444:doi
1342:896
1338:RFC
1091:or
1014:or
1008:TCP
988:UDP
982:BSD
978:TCP
645:max
285:by
165:by
115:TCP
97:in
74:as
35:in
2006::
1871:.
1852:.
1798:,
1719:.
1697:.
1683:14
1681:.
1677:.
1656:,
1646:,
1607:^
1556:.
1552:,
1532:,
1522:,
1482:22
1480:.
1476:.
1450:.
1440:33
1434:.
1408:.
1389:,
1234:.
1183:3G
1181:,
1064:.
1045:.
740:,
536:,
399:,
247:.
239:,
181:.
149:.
133:.
109:,
78:.
63:.
51:,
1927:)
1893:.
1881:.
1862:.
1843:.
1792::
1729:.
1705:.
1693::
1518::
1488::
1458:.
1446::
1419:.
1369:.
871:l
849:l
845:p
821:,
816:i
813:l
809:r
803:l
799:p
793:l
785:=
780:i
776:y
753:l
749:p
713:c
707:x
704:R
680:)
675:i
671:x
667:(
664:U
659:i
649:x
621:x
593:)
590:x
587:(
584:U
564:R
544:c
524:x
504:l
484:i
462:i
459:l
455:r
434:l
412:l
408:c
387:i
365:i
361:x
312:)
306:(
301:)
297:(
293:.
279:.
20:)
Text is available under the Creative Commons Attribution-ShareAlike License. Additional terms may apply.