As per Relevance of the word congestion, we have this rfc below:
Network Working Group B. Braden, USC/
Request for Comments: 2309 D. Clark, MIT
Category: Informational J. Crowcroft,
B. Davie, Cisco
S. Deering, Cisco
D. Estrin,
S. Floyd,
V. Jacobson,
G. Minshall,
C. Partridge,
L. Peterson, University of
K. Ramakrishnan, ATT Labs
S. Shenker, Xerox
J. Wroclawski, MIT
L. Zhang,
April 1998
Recommendations on Queue Management and Congestion
in the
Status of
This memo provides information for the Internet community.
does not specify an Internet standard of any kind.
of this memo is unlimited
Copyright
Copyright (C) The Internet Society (1998). All Rights Reserved
This memo presents two recommendations to the Internet
concerning measures to improve and preserve Internet performance
It presents a strong recommendation for testing, standardization
and widespread deployment of active queue management in routers
to improve the performance of today's Internet. It also urges
concerted effort of research, measurement, and ultimate
of router mechanisms to protect the Internet from flows that
not sufficiently responsive to congestion notification
Braden, et. al. Informational [Page 1]
RFC 2309 Internet Performance Recommendations April 1998
1.
The Internet protocol architecture is based on a connectionless end
to-end packet service using the IP protocol. The advantages of
connectionless design, flexibility and robustness, have been
demonstrated. However, these advantages are not without cost
careful design is required to provide good service under heavy load
In fact, lack of attention to the dynamics of packet forwarding
result in severe service degradation or "Internet meltdown".
phenomenon was first observed during the early growth phase of
Internet of the mid 1980s [Nagle84], and is technically
"congestion collapse".
The original fix for Internet meltdown was provided by Van Jacobson
Beginning in 1986, Jacobson developed the congestion
mechanisms that are now required in TCP implementations [Jacobson88,
HostReq89]. These mechanisms operate in the hosts to cause
connections to "back off" during congestion. We say that TCP
are "responsive" to congestion signals (i.e., dropped packets)
the network. It is primarily these TCP congestion
algorithms that prevent the congestion collapse of today's Internet
However, that is not the end of the story. Considerable research
been done on Internet dynamics since 1988, and the Internet
grown. It has become clear that the TCP congestion
mechanisms [RFC2001], while necessary and powerful, are
sufficient to provide good service in all circumstances. Basically
there is a limit to how much control can be accomplished from
edges of the network. Some mechanisms are needed in the routers
complement the endpoint congestion avoidance mechanisms
It is useful to distinguish between two classes of router
related to congestion control: "queue management" versus "scheduling
algorithms. To a rough approximation, queue management
manage the length of packet queues by dropping packets when
or appropriate, while scheduling algorithms determine which packet
send next and are used primarily to manage the allocation
bandwidth among flows. While these two router mechanisms are
related, they address rather different performance issues
This memo highlights two router performance issues. The first
is the need for an advanced form of router queue management that
call "active queue management." Section 2 summarizes the
that active queue management can bring. Section 3 describes
recommended active queue management mechanism, called Random
Detection or "RED". We expect that the RED algorithm can be
with a wide variety of scheduling algorithms, can be
relatively efficiently, and will provide significant
Braden, et. al. Informational [Page 2]
RFC 2309 Internet Performance Recommendations April 1998
performance improvement
The second issue, discussed in Section 4 of this memo, is
potential for future congestion collapse of the Internet due to
that are unresponsive, or not sufficiently responsive, to
indications. Unfortunately, there is no consensus solution
controlling congestion caused by such aggressive flows;
research and engineering will be required before any solution will
available. It is imperative that this work be energetically pursued
to ensure the future stability of the Internet
Section 5 concludes the memo with a set of recommendations to
IETF concerning these topics
The discussion in this memo applies to "best-effort" traffic.
Internet integrated services architecture, which provides a
for protecting individual flows from congestion, introduces its
queue management and scheduling algorithms [Shenker96, Wroclawski96].
Similarly, the discussion of queue management and congestion
requirements for differential services is a separate issue. However
we do not expect the deployment of integrated services
differential services to significantly diminish the importance of
best-effort traffic issues discussed in this memo
Preparation of this memo resulted from past discussions of end-to-
performance, Internet congestion, and RED in the End-to-End
Group of the Internet Research Task Force (IRTF).
2. THE NEED FOR ACTIVE QUEUE
The traditional technique for managing router queue lengths is to
a maximum length (in terms of packets) for each queue, accept
for the queue until the maximum length is reached, then reject (drop
subsequent incoming packets until the queue decreases because
packet from the queue has been transmitted. This technique is
as "tail drop", since the packet that arrived most recently (i.e.,
the one on the tail of the queue) is dropped when the queue is full
This method has served the Internet well for years, but it has
important drawbacks
1. Lock-
In some situations tail drop allows a single connection or a
flows to monopolize queue space, preventing other
from getting room in the queue. This "lock-out" phenomenon
often the result of synchronization or other timing effects
Braden, et. al. Informational [Page 3]
RFC 2309 Internet Performance Recommendations April 1998
2. Full
The tail drop discipline allows queues to maintain a full (or
almost full) status for long periods of time, since tail
signals congestion (via a packet drop) only when the queue
become full. It is important to reduce the steady-state
size, and this is perhaps queue management's most
goal
The naive assumption might be that there is a simple
between delay and throughput, and that the recommendation
queues be maintained in a "non-full" state
translates to a recommendation that low end-to-end delay is
important than high throughput. However, this does not
into account the critical role that packet bursts play
Internet performance. Even though TCP constrains a flow'
window size, packets often arrive at routers in
[Leland94]. If the queue is full or almost full, an
burst will cause multiple packets to be dropped. This
result in a global synchronization of flows throttling back
followed by a sustained period of lowered link utilization
reducing overall throughput
The point of buffering in the network is to absorb data
and to transmit them during the (hopefully) ensuing bursts
silence. This is essential to permit the transmission of
data. It should be clear why we would like to have normally
small queues in routers: we want to have queue capacity
absorb the bursts. The counter-intuitive result is
maintaining normally-small queues can result in
throughput as well as lower end-to-end delay. In short,
limits should not reflect the steady state queues we
maintained in the network; instead, they should reflect the
of bursts we need to absorb
Besides tail drop, two alternative queue disciplines that can
applied when the queue becomes full are "random drop on full"
"drop front on full". Under the random drop on full discipline,
router drops a randomly selected packet from the queue (which can
an expensive operation, since it naively requires an O(N)
through the packet queue) when the queue is full and a new
arrives. Under the "drop front on full" discipline [Lakshman96],
router drops the packet at the front of the queue when the queue
full and a new packet arrives. Both of these solve the lock-
problem, but neither solves the full-queues problem described above
Braden, et. al. Informational [Page 4]
RFC 2309 Internet Performance Recommendations April 1998
We know in general how to solve the full-queues problem
"responsive" flows, i.e., those flows that throttle back in
to congestion notification. In the current Internet, dropped
serve as a critical mechanism of congestion notification to
nodes. The solution to the full-queues problem is for routers
drop packets before a queue becomes full, so that end nodes
respond to congestion before buffers overflow. We call such
proactive approach "active queue management". By dropping
before buffers overflow, active queue management allows routers
control when and how many packets to drop. The next
introduces RED, an active queue management mechanism that solves
problems listed above (given responsive flows).
In summary, an active queue management mechanism can provide
following advantages for responsive flows
1. Reduce number of packets dropped in
Packet bursts are an unavoidable aspect of packet
[Willinger95]. If all the queue space in a router is
committed to "steady state" traffic or if the buffer space
inadequate, then the router will have no ability to
bursts. By keeping the average queue size small, active
management will provide greater capacity to absorb naturally
occurring bursts without dropping packets
Furthermore, without active queue management, more packets
be dropped when a queue does overflow. This is undesirable
several reasons. First, with a shared queue and the tail
discipline, an unnecessary global synchronization of
cutting back can result in lowered average link utilization,
hence lowered network throughput. Second, TCP recovers
more difficulty from a burst of packet drops than from a
packet drop. Third, unnecessary packet drops represent
possible waste of bandwidth on the way to the drop point
We note that while RED can manage queue lengths and reduce end
to-end latency even in the absence of end-to-end
control, RED will be able to reduce packet dropping only in
environment that continues to be dominated by end-to-
congestion control
2. Provide lower-delay interactive
By keeping the average queue size small, queue management
reduce the delays seen by flows. This is particularly
for interactive applications such as short Web transfers,
traffic, or interactive audio-video sessions, whose
Braden, et. al. Informational [Page 5]
RFC 2309 Internet Performance Recommendations April 1998
(and objective) performance is better when the end-to-end
is low
3. Avoid lock-out
Active queue management can prevent lock-out behavior
ensuring that there will almost always be a buffer available
an incoming packet. For the same reason, active
management can prevent a router bias against low bandwidth
highly bursty flows
It is clear that lock-out is undesirable because it
a gross unfairness among groups of flows. However, we
short of calling this benefit "increased fairness",
general fairness among flows requires per-flow state, which
not provided by queue management. For example, in a
using queue management but only FIFO scheduling, two TCP
may receive very different bandwidths simply because they
different round-trip times [Floyd91], and a flow that does
use congestion control may receive more bandwidth than a
that does. Per-flow state to achieve general fairness might
maintained by a per-flow scheduling algorithm such as
Queueing (FQ) [Demers90], or a class-based scheduling
such as CBQ [Floyd95], for example
On the other hand, active queue management is needed even
routers that use per-flow scheduling algorithms such as FQ
class-based scheduling algorithms such as CBQ. This is
per-flow scheduling algorithms by themselves do nothing
control the overall queue size or the size of individual queues
Active queue management is needed to control the overall
queue sizes, so that arriving bursts can be accommodated
dropping packets. In addition, active queue management
be used to control the queue size for each individual flow
class, so that they do not experience unnecessarily high delays
Therefore, active queue management should be applied across
classes or flows as well as within each class or flow
In short, scheduling algorithms and queue management should
seen as complementary, not as replacements for each other.
particular, there have been implementations of queue
added to FQ, and work is in progress to add RED queue
to CBQ
Braden, et. al. Informational [Page 6]
RFC 2309 Internet Performance Recommendations April 1998
3. THE QUEUE MANAGEMENT ALGORITHM "RED
Random Early Detection, or RED, is an active queue
algorithm for routers that will provide the Internet
advantages cited in the previous section [RED93]. In contrast
traditional queue management algorithms, which drop packets only
the buffer is full, the RED algorithm drops arriving
probabilistically. The probability of drop increases as
estimated average queue size grows. Note that RED responds to
time-averaged queue length, not an instantaneous one. Thus, if
queue has been mostly empty in the "recent past", RED won't tend
drop packets (unless the queue overflows, of course!). On the
hand, if the queue has recently been relatively full,
persistent congestion, newly arriving packets are more likely to
dropped
The RED algorithm itself consists of two main parts: estimation
the average queue size and the decision of whether or not to drop
incoming packet
(a) Estimation of Average Queue
RED estimates the average queue size, either in the
path using a simple exponentially weighted moving average (
as presented in Appendix A of [Jacobson88]), or in
background (i.e., not in the forwarding path) using a
mechanism
Note: The queue size can be measured either in units
packets or of bytes. This issue is discussed briefly
[RED93] in the "Future Work" section
Note: when the average queue size is computed in
forwarding path, there is a special case when a
arrives and the queue is empty. In this case,
computation of the average queue size must take into
how much time has passed since the queue went empty. This
discussed further in [RED93].
(b) Packet Drop
In the second portion of the algorithm, RED decides whether
not to drop an incoming packet. It is RED's
algorithm for dropping that results in performance
for responsive flows. Two RED parameters, minth (
threshold) and maxth (maximum threshold), figure prominently
Braden, et. al. Informational [Page 7]
RFC 2309 Internet Performance Recommendations April 1998
this decision process. Minth specifies the average queue
*below which* no packets will be dropped, while maxth
the average queue size *above which* all packets will
dropped. As the average queue size varies from minth to maxth
packets will be dropped with a probability that varies
from 0 to maxp
Note: a simplistic method of implementing this would be
calculate a new random number at each packet arrival,
compare that number with the above probability which
from 0 to maxp. A more efficient implementation,
in [RED93], computes a random number *once* for each
packet
Note: the decision whether or not to drop an incoming
can be made in "packet mode", ignoring packet sizes, or
"byte mode", taking into account the size of the
packet. The performance implications of the choice
packet mode or byte mode is discussed further in [Floyd97].
RED effectively controls the average queue size while
accommodating bursts of packets without loss. RED's use
randomness breaks up synchronized processes that lead to lock-
phenomena
There have been several implementations of RED in routers, and
have been published reporting on experience with
implementations ([Villamizar94], [Gaynor96]). Additional reports
implementation experience would be welcome, and will be posted on
RED web page [REDWWW].
All available empirical evidence shows that the deployment of
queue management mechanisms in the Internet would have
performance benefits. There are seemingly no disadvantages to
the RED algorithm, and numerous advantages. Consequently, we
that the RED active queue management algorithm should be
deployed
We should note that there are some extreme scenarios for which
will not be a cure, although it won't hurt and may still help.
example of such a scenario would be a very large number of flows
each so tiny that its fair share would be less than a single
per RTT
Braden, et. al. Informational [Page 8]
RFC 2309 Internet Performance Recommendations April 1998
4. MANAGING AGGRESSIVE
One of the keys to the success of the Internet has been
congestion avoidance mechanisms of TCP. Because TCP "backs off
during congestion, a large number of TCP connections can share
single, congested link in such a way that bandwidth is
reasonably equitably among similarly situated flows. The
sharing of bandwidth among flows depends on the fact that all
are running basically the same congestion avoidance algorithms
conformant with the current TCP specification [HostReq89].
We introduce the term "TCP-compatible" for a flow that behaves
congestion like a flow produced by a conformant TCP. A TCP
compatible flow is responsive to congestion notification, and
steady-state it uses no more bandwidth than a conformant TCP
under comparable conditions (drop rate, RTT, MTU, etc.)
It is convenient to divide flows into three classes: (1) TCP
compatible flows, (2) unresponsive flows, i.e., flows that do
slow down when congestion occurs, and (3) flows that are
but are not TCP-compatible. The last two classes contain
aggressive flows that pose significant threats to
performance, as we will now discuss
o Non-Responsive
There is a growing set of UDP-based applications
congestion avoidance algorithms are inadequate or
(i.e, the flow does not throttle back upon receipt of
notification). Such UDP applications include
applications like packet voice and video, and also
bulk data transport [SRM96]. If no action is taken,
unresponsive flows could lead to a new congestion collapse
In general, all UDP-based streaming applications
incorporate effective congestion avoidance mechanisms.
example, recent research has shown the possibility
incorporating congestion avoidance mechanisms such as Receiver
driven Layered Multicast (RLM) within UDP-based
applications such as packet video [McCanne96; Bolot94].
research and development on ways to accomplish
avoidance for streaming applications will be very important
However, it will also be important for the network to be able
protect itself against unresponsive flows, and mechanisms
accomplish this must be developed and deployed. Deployment
such mechanisms would provide incentive for every
application to become responsive by incorporating its
Braden, et. al. Informational [Page 9]
RFC 2309 Internet Performance Recommendations April 1998
congestion control
o Non-TCP-Compatible Transport
The second threat is posed by transport protocol
that are responsive to congestion notification but,
deliberately or through faulty implementations, are not TCP
compatible. Such applications can grab an unfair share of
network bandwidth
For example, the popularity of the Internet has caused
proliferation in the number of TCP implementations. Some
these may fail to implement the TCP congestion
mechanisms correctly because of poor implementation. Others
deliberately be implemented with congestion avoidance
that are more aggressive in their use of bandwidth than
TCP implementations; this would allow a vendor to claim to
a "faster TCP". The logical consequence of such
would be a spiral of increasingly aggressive
implementations, leading back to the point where there
effectively no congestion avoidance and the Internet
chronically congested
Note that there is a well-known way to achieve more
TCP performance without even changing TCP: open
connections to the same place, as has been done in some
browsers
The projected increase in more aggressive flows of both
classes, as a fraction of total Internet traffic, clearly poses
threat to the future Internet. There is an urgent need
measurements of current conditions and for further research into
various ways of managing such flows. There are many difficult
in identifying and isolating unresponsive or non-TCP-compatible
at an acceptable router overhead cost. Finally, there is
measurement or simulation evidence available about the rate at
these threats are likely to be realized, or about the
benefit of router algorithms for managing such flows
There is an issue about the appropriate granularity of a "flow".
There are a few "natural" answers: 1) a TCP or UDP connection (
address/port, destination address/port); 2) a source/destination
pair; 3) a given source host or a given destination host. We
guess that the source/destination host pair gives the
appropriate granularity in many circumstances. However, it
possible that different vendors/providers could set
granularities for defining a flow (as a way of "distinguishing
themselves from one another), or that different granularities
Braden, et. al. Informational [Page 10]
RFC 2309 Internet Performance Recommendations April 1998
be chosen for different places in the network. It may be the
that the granularity is less important than the fact that we
dealing with more unresponsive flows at *some* granularity.
granularity of flows for congestion management is, at least in part
a policy question that needs to be addressed in the wider
community
5. CONCLUSIONS AND
This discussion leads us to make the following recommendations to
IETF and to the Internet community as a whole
o RECOMMENDATION 1:
Internet routers should implement some active queue
mechanism to manage queue lengths, reduce end-to-end latency
reduce packet dropping, and avoid lock-out phenomena within
Internet
The default mechanism for managing queue lengths to meet
goals in FIFO queues is Random Early Detection (RED) [RED93].
Unless a developer has reasons to provide another
mechanism, we recommend that RED be used
o RECOMMENDATION 2:
It is urgent to begin or continue research, engineering,
measurement efforts contributing to the design of mechanisms
deal with flows that are unresponsive to congestion
or are responsive but more aggressive than TCP
Although there has already been some limited deployment of RED in
Internet, we may expect that widespread implementation and
of RED in accordance with Recommendation 1 will expose a number
engineering issues. For example, such issues may include
implementation questions for Gigabit routers, the use of RED in
2 switches, and the possible use of additional considerations,
as priority, in deciding which packets to drop
We again emphasize that the widespread implementation and
of RED would not, in and of itself, achieve the goals
Recommendation 2.
Widespread implementation and deployment of RED will also enable
introduction of other new functionality into the Internet.
example of an enabled functionality would be the addition of
congestion notification [Ramakrishnan97] to the
architecture, as a mechanism for congestion notification in
Braden, et. al. Informational [Page 11]
RFC 2309 Internet Performance Recommendations April 1998
to packet drops. A second example of new functionality would
implementation of queues with packets of different drop priorities
packets would be transmitted in the order in which they arrived,
during times of congestion packets of the lower drop priority
be preferentially dropped
6.
[Bolot94] Bolot, J.-C., Turletti, T., and Wakeman, I.,
Feedback Control for Multicast Video Distribution in the Internet
ACM SIGCOMM '94, Sept. 1994.
[Demers90] Demers, A., Keshav, S., and Shenker, S., Analysis
Simulation of a Fair Queueing Algorithm, Internetworking:
and Experience, Vol. 1, 1990, pp. 3-26.
[Floyd91] Floyd, S., Connections with Multiple Congested Gateways
Packet-Switched Networks Part 1: One-way Traffic.
Communications Review, Vol.21, No.5, October 1991, pp. 30-47.
http://ftp.ee.lbl.gov/floyd/.
[Floyd95] Floyd, S., and Jacobson, V., Link-sharing and
Management Models for Packet Networks. IEEE/ACM Transactions
Networking, Vol. 3 No. 4, pp. 365-386, August 1995.
[Floyd97] Floyd, S., RED: Discussions of Byte and Packet Modes,
1997 email, http://www-nrg.ee.lbl.gov/floyd/REDaveraging.txt
[Gaynor96] Gaynor, M., Proactive Packet Dropping Methods for
Gateways, October 1996, URL http://www.eecs.harvard.edu/~gaynor
final.ps
[HostReq89] Braden, R., Ed., "Requirements for Internet Hosts --
Communication Layers", STD 3, RFC 1122, October 1989.
[Jacobson88] V. Jacobson, Congestion Avoidance and Control,
SIGCOMM '88, August 1988.
[Lakshman96] T. V. Lakshman, Arnie Neidhardt, Teunis Ott, The
From Front Strategy in TCP Over ATM and Its Interworking with
Control Features, Infocom 96, MA28.1.
[Leland94] W. Leland, M. Taqqu, W. Willinger, and D. Wilson, On
Self-Similar Nature of Ethernet Traffic (Extended Version), IEEE/
Transactions on Networking, 2(1), pp. 1-15, February 1994.
Braden, et. al. Informational [Page 12]
RFC 2309 Internet Performance Recommendations April 1998
[McCanne96] McCanne, S., Jacobson, V., and M. Vetterli, Receiver
driven Layered Multicast, ACM
[Nagle84] Nagle, J., "Congestion Control in IP/TCP", RFC 896,
1984.
[Ramakrishnan97] Ramakrishnan, K. K., and S. Floyd, "A Proposal
add Explicit Congestion Notification (ECN) to IPv6 and to TCP",
in Progress
[RED93] Floyd, S., and Jacobson, V., Random Early Detection
for Congestion Avoidance, IEEE/ACM Transactions on Networking, V.1
N.4, August 1993, pp. 397-413. Also available
http://ftp.ee.lbl.gov/floyd/red.html
[REDWWW] Floyd, S., The RED Web Page, 1997,
http://ftp.ee.lbl.gov/floyd/red.html
[RFC 2001] Stevens, W., "TCP Slow Start, Congestion Avoidance,
Retransmit, and Fast Recovery Algorithms", RFC 2001, January 1997.
[Shenker96] Shenker, S., Partridge, C., and R. Guerin, "
of Guaranteed Quality of Service", Work in Progress
[SRM96] Floyd. S., Jacobson, V., McCanne, S., Liu, C., and L. Zhang
A Reliable Multicast Framework for Light-weight Sessions
Application Level Framing. ACM SIGCOMM '96, pp 342-355.
[Villamizar94] Villamizar, C., and Song, C., High Performance TCP
ANSNET. Computer Communications Review, V. 24 N. 5, October 1994, pp
45-60. URL http://ftp.ans.net/pub/papers/tcp-performance.ps
[Willinger95] W. Willinger, M. S. Taqqu, R. Sherman, D. V. Wilson
Self-Similarity Through High-Variability: Statistical Analysis
Ethernet LAN Traffic at the Source Level, ACM SIGCOMM '95, pp. 100-
113, August 1995.
[Wroclawski96] Wroclawski, J., "Specification of the Controlled-
Network Element Service", Work in Progress
Braden, et. al. Informational [Page 13]
RFC 2309 Internet Performance Recommendations April 1998
Security
While security is a very important issue, it is largely orthogonal
the performance issues discussed in this memo. We note, however
that denial-of-service attacks may create unresponsive traffic
that are indistinguishable from flows from normal high-
isochronous applications, and the mechanism suggested
Recommendation 2 will be equally applicable to such attacks
Authors'
Bob
USC Information Sciences
4676 Admiralty
Marina del Rey, CA 90292
Phone: 310-822-1511
EMail: Braden@ISI.
David D.
MIT Laboratory for Computer
545 Technology Sq
Cambridge, MA 02139
Phone: 617-253-6003
EMail: DDC@lcs.mit.
Jon
University College
Department of Computer
Gower
London, WC1E 6
Phone: +44 171 380 7296
EMail: Jon.Crowcroft@cs.ucl.ac.
Bruce
Cisco Systems, Inc
250 Apollo
Chelmsford, MA 01824
Phone
EMail: bdavie@cisco.
Braden, et. al. Informational [Page 14]
RFC 2309 Internet Performance Recommendations April 1998
Steve
Cisco Systems, Inc
170 West Tasman
San Jose, CA 95134-1706
Phone: 408-527-8213
EMail: deering@cisco.
Deborah
USC Information Sciences
4676 Admiralty
Marina del Rey, CA 90292
Phone: 310-822-1511
EMail: Estrin@usc.
Sally
Lawrence Berkeley National Laboratory
MS 50B-2239,
One Cyclotron Road
Berkeley CA 94720
Phone: 510-486-7518
EMail: Floyd@ee.lbl.
Van
Lawrence Berkeley National Laboratory
MS 46A
One Cyclotron Road
Berkeley CA 94720
Phone: 510-486-7519
EMail: Van@ee.lbl.
Greg
Fiberlane
1399 Charleston
Mountain View, CA 94043
Phone: +1 650 237 3164
EMail: Minshall@fiberlane.
Braden, et. al. Informational [Page 15]
RFC 2309 Internet Performance Recommendations April 1998
Craig
BBN
10 Moulton St
Cambridge MA 02138
Phone: 510-558-8675
EMail: craig@bbn.
Larry
Department of Computer
University of
Tucson, AZ 85721
Phone: 520-621-4231
EMail: LLP@cs.arizona.
K. K.
AT&T Labs.
Rm. A155
180 Park
Florham Park, N.J. 07932
Phone: 973-360-8766
EMail: KKRama@research.att.
Scott
Xerox
3333 Coyote Hill
Palo Alto, CA 94304
Phone: 415-812-4840
EMail: Shenker@parc.xerox.
John
MIT Laboratory for Computer
545 Technology Sq
Cambridge, MA 02139
Phone: 617-253-7885
EMail: JTW@lcs.mit.
Lixia
4531G Boelter
Los Angeles, CA 90024
Phone: 310-825-2695
EMail: Lixia@cs.ucla.
Braden, et. al. Informational [Page 16]
RFC 2309 Internet Performance Recommendations April 1998
Full Copyright
Copyright (C) The Internet Society (1998). All Rights Reserved
This document and translations of it may be copied and furnished
others, and derivative works that comment on or otherwise explain
or assist in its implementation may be prepared, copied,
and distributed, in whole or in part, without restriction of
kind, provided that the above copyright notice and this paragraph
included on all such copies and derivative works. However,
document itself may not be modified in any way, such as by
the copyright notice or references to the Internet Society or
Internet organizations, except as needed for the purpose
developing Internet standards in which case the procedures
copyrights defined in the Internet Standards process must
followed, or as required to translate it into languages other
English
The limited permissions granted above are perpetual and will not
revoked by the Internet Society or its successors or assigns
This document and the information contained herein is provided on
"AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET
TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED,
BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE
HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES
MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE
Braden, et. al. Informational [Page 17]
if you see any problems within the linking, don't worry be happy,
this is version 0.1 of the Relevance System and you gotta expect some crappy subroutines sometimes,
just be content we did not write this in Java, which would have made this "bigger and better" HAHAHHA.
RFC documents can be found at I.E.T.F.
Relevance System Copyright © 2002 Spectrum WorldResearch
other technical nosh by ServerMasters Corporation
collaboration of BobX
