Ramakrishnan, K.K., and Jain, R., "A Binary Feedback Scheme for Congestion Avoidance in Computer Networks", ACM Transactions on Computer Systems, Vol. 8 No. 2, pp. 158-181, 1990.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....divided into two groups, according to the type of feedback they supply: binary (one bit) feedback and explicit rate feedback. In binary feedback algorithms, the source periodically receives a single bit, on a control message (or in some cases, piggybacked on backward direction data packet as in [29]) which indicates whether there is congestion in the network. Two of the earlier proposals to the ATM Forum were the Forward Explicit Congestion Notification (FECN) 40] scheme and the Explicit Congestion Notification (BECN) 27] scheme. In the FECN scheme, each data cell holds an Explicit ....

K. Ramakrishnan and R. Jain. A binary feedback scheme for congestion avoidance in computer networks. IEEE Transactions on Communications, 8(2):158--181, May 1990.

....and the resulting low service rates. With the advent of packet networks, there appeared a possibility to control the arrival rate by notifying the senders. The mechanism used by TCP [13] is to take packet loss as an indication. A less drastic measure was later used in the DECbit feedback mechanism [14], which used a bit in the DECnet protocol header but required symmetric routes. ECN [6] recently brought feedback into the Internet world while avoiding the symmetric route requirement, which significantly reduced the requirements for retransmits [15] DECbit already provided a first early ....

K. K. Ramakrishnan and Raj Jain. A binary feedback scheme for congestion avoidance in computer networks. ACM Transactions on Computer Systems, 8(2):158--181, 1990.

....and widely used active queue management algorithm. Recently an alternative algorithm called BLUE [7] is developed. In this section these two active queue management algorithms are briefly described. 2.1. RED RED, a development on the previous works like ERD (Early Random Drop) 8] and DECbit [9], is the first widely used active queue management algorithm. ERD drops each arriving packet at the gateway with a fixed drop probability once the queue length exceeds a certain drop level. It is shown that misbehaving users receives higher throughput with ERD [10] DECbit is a binary feedback ....

Ramakrishnan, K.K., and Jain, R., 1990, A binary feedback scheme for congestion avoidance in computer networks, ACM Transactions on Computer Systems, V.8, N.2, pp. 152-181

....packet and therefore global prioritization across all flows from multiple servers can be performed at an intermediate resource bottleneck within the network. Information exchange between data senders and routers can improve end to end data transmission. In Explicit Congestion Notification (ECN) [10], routers set the Congestion Experienced (CE) bit in the IP packet header to inform data senders of network congestion. In PBE STP, the information flows in the opposite direction from data senders to routers. 3. PBE STP: OVERVIEW AND DESIGN The general framework of the proposed system is ....

K.K. Ramakrishnan and R. Jain, "A binary feedback scheme for congestion avoidance in computer networks," ACM Transactions on Computer Systems (TOCS), vol. 8, no. 2, pp. 158--181, May 1990.

.... that the growing demand for network bandwidth has driven loss rates up across various links in the Internet [23] In order to stem the increasing packet loss rates caused by an exponential increase in network traffic, the IETF is considering the deployment of explicit congestion notification (ECN) [11, 24, 25] along with active queue management techniques such as RED (Random Early Detection) 2, 11] While ECN is necessary for eliminating packet loss in the Internet [10] this paper shows that RED, even when used in conjunction with ECN, is ineffective in preventing packet loss. The basic idea behind ....

K. K. Ramakrishan and R. Jain. A Binary Feedback Scheme for Congestion Avoidance in Computer Networks. ACM Transaction on Computer Systems, 8(2):158--181, May 1990.

.... that the growing demand for network bandwidth has driven loss rates up across various links in the Internet [28] In order to stem the increasing packet loss rates caused by an exponential increase in network traffic, the IETF is considering the deployment of explicit congestion notification (ECN) [12, 30, 31] along with active queue management techniques such as RED (Random Early Detection) 3, 14] While ECN is necessary for eliminating packet loss in the Internet [9] we show that RED, even when used in conjunction with ECN, is ineffective in preventing packet loss. The basic idea behind RED queue ....

K. K. Ramakrishan and R. Jain. A Binary Feedback Scheme for Congestion Avoidance in Computer Networks. ACM Transaction on Computer Systems, 8(2):158--181, May 1990.

....congestion to end nodes. One is to use packet drops. Alternatively, it can set a Congestion Experienced (CE) bit in a packet header as an indication of congestion, instead of relying solely on packet drops. The latter method is commonly referred to as explicit congestion notification (ECN) [17, 42, 43]. The major advantage of active queue management mechanisms like RED is that the transport protocols with congestion control (e.g. TCP) do not have to rely on buffer overflow as the only indication of congestion. This can potentially reduce unnecessary queuing delay for all traffic sharing that ....

....weighted moving average: q (1 w q ) q w q q, where w q is a fixed (small) parameter and q is the instantaneous queue length. When the average queue length exceeds a minimum threshold (min th ) incoming packets are probabilistically dropped or marked with the Congestion Experienced bit [17, 42, 43]. The probability that a packet arriving at the RED queue is either dropped or marked depends upon several control parameters of the algorithm. An initial drop mark probability P b is computed using a drop function D based on the average queue length q and three control parameters max p , min th ....

K. Ramakrishnan and R. Jain. A Binary Feedback Scheme for Congestion Avoidance in Computer Networks. ACM Transaction on Computer Systems, 8(2):158-181, May 1990.

.... period of time required for information to be communicated from source to destination and back to source, is a useful timescale to categorise mechanisms such as TCP ( Postel81b, Nagle85, Lottor88, Stevens97, Mathis96] and explicit congestion notification schemes proposed as alternatives to it [Ramakrishnan90, Floyd94, Ramakrishnan99, Key99, Gibbens99] Section 2.3.4 summarises the significant quantity of literature that exists describing control techniques at the call level, the timescale over which flows of data are connected. A call example would be a connection in the telephone network. Calls ....

K. K. Ramakrishnan and R. Jain. A Binary Feedback Scheme for Congestion Avoidance in Computer Networks. ACM Transactions on Computer Systems, 8(2):158--181, May 1990. (p 49)

....the source end system decreases its cell transmission rate after it receives congestion indication. Our main objective of the current paper is to evaluate the rate based flow control scheme by an analytic approach. Several analytic studies for the rate control scheme have already been made [1, 2, 3]. For example, Yin et al. analyzed a dynamical behavior based on the timer based approach [3] In their model, the source end system changes its cell transmission rate regularly at every fixed time interval. However, it has already been found that this approach causes problems in some situations ....

K.K.Ramakrishnan and R.Jain, "A binary feedback scheme for congestion avoidance in computer networks, " ACM Transactions on Computer Systems, vol. 8, no. 2, pp. 158--181, 1990.

.... to fair states, an LIMD algorithm should have an additive increase policy and a multiplicative decrease policy: aI 0; bI = 1; 0 bD 1: 7) This claim is important because its pick of AIMD from the stable class of LIMD algorithms justi es the use of AIMD for congestion control in DECbit [13] and TCP [1, 5] Besides, Assertion 2 has accidentally sprung the popular belief that AIMD is necessary for stability of congestion control [12] According to Assertion 1, this belief is incorrect: all LIMD algorithms including those that have a multiplicative component in their increase policy ....

K.K. Ramakrishnan and R. Jain. A Binary Feedback Scheme for Congestion Avoidance in Computer Networks with Connectionless Network Layer. In Proceedings ACM SIGCOMM'88, August 1988.

....results and live Internet based experiments are presented in Section VI and Section VII. Section VIII concludes the paper. II. RELATED WORK Many different approaches exist in the flow control literature, including TCP like window based protocols [6] 7] one or two bit feedback schemes [8], 9] 10] and optimization based flow control [11] 12] 13] 14] 15] 16] The economic framework for flow and congestion control used in many optimization based protocols[12] 14] has some similarity with the cost benefit model used in our work. In both, the links have some cost and ....

K. K. Ramakrishnan and R. Jain, "A binary feedback scheme for congestion avoidance in computer networks with a connectionless network layer," in Proceedings of ACM SIGCOMM, 1988.

....packets to a single address (a multicast address) and receives packets from a single address (due to Ack fusion) Thus, we can potentially have any existing unicast transport protocol at the sender. Specifically, we use TCP as the transport protocol. We can also use other schemes such as DECbit [RJ88] and RPCC [GV97] 4. Description of TCP M TCP M is scalable and easily deployable. It handles congestion in the network exactly like TCP. In the following subsections we describe how TCP M establishes and terminates a connection, how the data is transferred and Acks are fused and how network ....

K. K. Ramakrishnan and Raj Jain. A Binary Feedback Scheme for Congestion Avoidance in Computer Networks with a Connectionless Network Layer. In Proceedings of the ACM SIGCOMM '88 Symposium, pages 303-313, August 1988. Also available as DEC-TR-508, Digital Equipment Corporation.

....betterbest effort services, such as an end to end lossless TCP service illustrated in the paper. Our overlay architecture is unique in that it: a) does not require any support from intermediate network nodes or end systems (unlike ATM ABR [16] credit based schemes [20] bit based schemes [26], 7] or TCP modifications [13] 24] b) operates in a lossless (conditionally) manner with bounded buffers unlike TCP which uses packet loss to indicate congestion [12] c) aims to complement, not substitute existing end to end congestion control (eg: TCP) d) enables new edge based ....

....rate based concept because our goal is to match aggregate input rate to the bottleneck rate. Buffers are used merely to absorb statistical fluctuations and not considered part of the capacity of the network even if buffer space is unbounded. This is different from traditional window based [12] [26] or credit based [20] approaches which attempt to match aggregate window size to the bandwidth RTT product of the network where buffers are typically considered part of the bandwidth RTT product. Since our notion of congestion epoch implies bandwidth saturation and not buffer saturation, packet ....

[Article contains additional citation context not shown here]

RAMAKRISHNAN, K. K., AND JAIN, R. A binary feedback scheme for congestion avoidance in computer networks with a connectionless network layer. In Proceedings of ACM SIGCOMM '88 (August 1988), pp. 303--313.

....is familiar to teletraffic theorists, and our analytical approximations are taken from [16] In Section 2 we also outline some strategies by which packets may be marked, to indicate congestion. Packet marking strategies are familiar to flow control specialists, and are an active research area ([3,6,7,9,12,15]) In Section 3 we present a simple model of a gateway, and show that gateways are able to effect a distributed connection acceptance control for the network. We compare their performance with the performance achieved by a centralized controller, roughly summarized by Erlang s formula. In Section ....

K. K. Ramakrishnan and R. Jain. A binary feedback scheme for congestion avoidance in computer networks. ACM Trans. Computer Systems, 8:158-181, 1990.

....72 hours of data. 4. 1 Why Motivation is the distribution of delay important We believe there are several reasons: Distribution of end to end delay is an important component for modeling studies of feedback based congestion control where feedback delay is known to be an im portant component [2, 11, 16, 30, 33, 40] in performance. Since this distribution has hitherto been unknown, it has usually been assumed deterministic [2, 11, 30] essentially for lack of anything better. Knowledge of the actual delay distribution could make the models more accurate, show its impact on performance, and enable ....

.... to a base level congestion along a path that may be effectively tracked with a feedback mechanism If this were so, then even in high speed networks, the congestion information could be useful (in spite of delay in the feedback path) Most feedback based schemes proposed in the literature, e.g. [9, 16, 20, 22, 25, 27, 32, 33, 34], would benefit if low frequency components of delay were to show a correlation with packet loss. For example, the Decbit strategy [33] of averaging the number of congestion signals over an interval and adjusting the window for the following interval assumed a correlation of trattic across ....

[Article contains additional citation context not shown here]

Ramakrishnan, K.K., and R. Jain, "A Binary Feedback Scheme for Congestion Avoidance in Computer Networks with a Connectionless Network Layer," ACM Trans. Computer Systems, vol. 8, pp 158-181, May 1990.

....2 Congestion avoidance and control In this section, we focus on end to end mechanisms. One of the most important characteristics of such a mechanism is its feedback: What information does the source obtain about the state of the intermediate nodes. In the binary feedback scheme used in DECNET [18], each data packet received by the destination has a bit indicating whether or not the packet encountered an intermediate node with average queue size greater than 1; this information is sent back to the source in acknowledgement packets. In the Internet TCP, the feedback consists of the arrival ....

....feedback information is used to regulate the flow of data from the sources into the network. In rate based mechanisms, the source rate is regulated by adjusting interpacket gaps. Examples are found in the protocols NETBLT [4] VMTP [2] and XTP [3] However, most protocols currently in operation [18, 11] use a window based scheme, rather than the rate based scheme. Here, a limit referred to as the window size is placed on the number of packets that can be outstanding at the source, but no constraint is placed on the rate at which these packets can be sent. For example, in the binary feedback ....

Ramakrishnan K. K., Jain R., A Binary Feedback Scheme for Congestion Avoidance in Computer Networks, ACM Transactions on Computer Systems, Vol. 8, No. 2, pp. 158-181, May 1990.

....2 source 4 LSP 1 LSP 2 LSP 4 source 3 LSP 3 Figure 2: The simplified network model. LSP i is the primary path for source i. Specifically, we investigate if and to which degree binary feedback schemes and rate control schemes, such as, additive increase multiplicative decrease (AIMD) [5, 6, 10, 11, 12, 19], can be used to achieve traffic engineering objectives. Recently, considerable effort has been invested into scalable mechanisms for providing differentiated services in the Internet. For example, in [7] Elwalid et al. presented a multi path adaptive traffic engineering mechanism, called MATE, ....

.... following corollary: Corollary 1 A routing matrix M is PPF optimal if it satisfies M(i; j) maxffl i Gamma B i ; 0g : 15) 4 Multipath AIMD Additive Increase Multiplicative Decrease (AIMD) feedback algorithms are used extensively for flow and congestion control in computer networks [5, 6, 11, 10, 12, 19] and are widely held to be both efficient and fair in allocating traffic to network paths. These algorithms adjust the transmission rate of a sender based on feedback from the network following an additive increase multiplicative decrease rule. If the network is free of congestion, the ....

K. K. Ramakrishnan and R. Jain. A Binary Feedback Scheme for Congestion Avoidance in Computer Networks. ACM Transactions on Computer Systems, 8(2):158--181, 1990.

No context found.

K. K. Ramakrishnan and R. Jain, A binary feedback scheme for congestion avoidance in computer networks," ACM Transactions on Computer Systems, Vol. 8, No. 2, May 1990, pp. 158-181.

No context found.

K. K. Ramakrishnan and R. Jain, A binary feedback scheme for congestion avoidance in computer networks," ACM Transactions on Computer Systems, Vol. 8, No. 2, May 1990, pp. 158-181.

No context found.

K. K. Ramakrishnan and R. Jain, A binary feedback scheme for congestion avoidance in computer networks, ACM Transactions on Computer Systems,Vol. 8, No. 2, pp. 158-181, May 1990.

No context found.

K. K. Ramakrishnan and R. Jain, A binary feedback scheme for congestion avoidance in computer networks, ACM Transactions on Computer Systems,Vol. 8, No. 2, pp. 158-181, May 1990.

No context found.

K. K. Ramakrishnan and R. Jain, "A binary feedback scheme for congestion avoidance in computer networks," ACM Transactions on Computer Systems, Vol. 8, No. 2, May 1990, pp. 158-181.

....send the packets in order. If some links have selective local link layer retransmission, like those used in wireless links to combat transmission errors, the packets are not delivered in order. The inference of congestion from duplicate ACKs us no longer valid. Ramakrishnan and Jain s work in [8], which has been popularly called the DECbit scheme, uses a single bit in the network layer header to signal the congestion. The Explicit Congestion Notification (ECN) 9, 10] motivated by the DECbit scheme, provides a mechanism for intermediate routers to send early congestion feedback to the ....

K. Ramakrishnan and R. Jain, A binary feedback scheme for congestion avoidance in computer networks , ACM Transactions on Computer Systems, Vol. 8, No. 2, pp. 158-181, May 1990.

....The packets are dropped when the queue exceeds the buffer capacity. The transmission control protocol (TCP) uses the packet drop as a signal of congestion and reduces its load [18] While in the past, this strategy has worked satisfactorily, now we need better strategies for two reasons [12, 15, 16]. First, the bandwidth of the networks as well as the distances are increasing. For very high distance bandwidth product networks, packet drop is not the optimal congestion indication. Several megabytes of data may be lost in the time required to detect and respond to packet losses. Therefore, a ....

K. K. Ramakrishnan and R. Jain, A binary feedback scheme for congestion avoidance in computer networks," ACM Transactions on Computer Systems, Vol. 8, No. 2, May 1990, pp. 158-181.

No context found.

Ramakrishnan, K.K., and Jain, R., "A Binary Feedback Scheme for Congestion Avoidance in Computer Networks", ACM Transactions on Computer Systems, Vol. 8 No. 2, pp. 158-181, 1990.

No context found.

K. K. Ramakrishnan and Raj Jain. A Binary Feedback Scheme for Congestion Avoidance in Computer Networks with a Connectionless Network Layer. In SIGCOMM 98 Symposium Proceedings on Communications Architectures and Protocols, pages 303--313, 1988.

No context found.

K. Ramakrishnan and R. Jain, "A Binary Feedback Scheme for Congestion Avoidance in Computer Networks," ACM Transactions on Computer Systems, vol. 8, no. 2, pp. 158--181, 1990.

No context found.

K. Ramakrishnan and R. Jain, "A Binary Feedback Scheme for Congestion Avoidance in Computer Networks with Connectionless Network Layer," Proc. SIGCOMM'88, August 1988.

No context found.

K. Ramakrishnan and R. Jain, A binary feedback scheme for congestion avoidance in computer networks with connectionless network layer, ACM SIGCOMM'88, August 1988.

No context found.

K. Ramakrishnan and R. Jain, "A Binary Feedback Scheme for Congestion Avoidance in Computer Networks with Connectionless Network Layer," Proc. SIGCOMM'88, August 1988.

No context found.

K. Ramakrishnan and R. Jain, "A Binary Feedback Scheme for Congestion Avoidance in Computer Networks with Connectionless Network Layer," Proc. SIGCOMM'88, August 1988.

No context found.

K. K. Ramakrishnan and Raj Jain. A binary feedback scheme for congestion avoidance in computer networks. ACM Transactions on Computer Systems, 8(2):158--181, May 1990.

No context found.

K. K. Ramakrishnan and R. Jain. A binary feedback scheme for congestion avoidance in computer networks with a connectionless network layer. Proceedings of SIGCOMM'88, ACM, August 1988.

No context found.

K. Ramakrishnan and R. Jain, "A binary feedback scheme for congestion avoidance in computer networks," ACM Trans. Comput. Syst., vol. 8, pp. 158-181, 1990.

No context found.

K.K. Ramakrishnan and R. Jain. A Binary Feedback Scheme for Congestion Avoidance in Computer Networks with Connectionless Network Layer. In Proceedings ACM SIGCOMM'88, August 1988.

No context found.

K. Ramakrishnan and R. Jain, "A binary feedback scheme for congestion avoidance in computer networks," ACM Trans. Comput. Syst., vol. 8, pp. 158-181, 1990.

No context found.

K. K. Ramakrishnan and Raj Jain. A binary feedback scheme for congestion avoidance in computer networks. ACM Transactions on Computer Systems, 8(2):158--181, 1990.

No context found.

K. K. Ramakrishnan and Raj Jain. A binary feedback scheme for congestion avoidance in computer networks. ACM Transactions on Computer Systems, 8(2):158-181, May 1990. 29

No context found.

K. Ramakrishnan and R. Jain. A binary feedback scheme for congestion avoidance in computer networks. ACM Trans. on Comm. Syst., 8(2), 1990, pp. 158--181.

No context found.

K. Ramakrishnan and R. Jain. A binary feedbackscheme for congestion avoidance in computer networks with connectionless network layer. ACM Transactions on Computers, 1990.

No context found.

K. K. Ramakrishnan and R. Jain. "A Binary feedback scheme for congestion avoidance in computer networks." ACM Trans. Comput. Syst.,8(2):158-181, 1990.

No context found.

K. K. Ramakrishnan and Raj Jain. A binary feedback scheme for congestion avoidance in computer networks. ACM Transactions on Computer Systems, 8(2):158--181, May 1990.

No context found.

K.K. Ramakrishnan and R. Jain. A binary feedback scheme for congestion avoidance in computer networks with connectionless network layer. In Proceedings ACM SIGCOMM'88, August 1988.

No context found.

K.K. Ramakrishnan and Raj Jain, "A Binary Feedback Scheme for Congestion Avoidance in Computer Networks," ACM Transactions on Computer Systems, vol.8, no.2, pp.158-181, May 1990.

No context found.

K. Ramakrishnan and R. Jain. A binary feedbackscheme for congestion avoidance in computer networks with connectionless network layer. ACM Transactions on Computers, 1990.

No context found.

K. Ramakrishnan and R. Jain. A binary feedback scheme for congestion avoidance in computer networks. ACM Trans. on Comm. Syst., 8(2), 1990, pp. 158--181.

No context found.

K. K. Ramakrishnan and Raj Jain. A Binary Feedback Scheme for Congestion Avoidance in Computer Networks with a Connectionless Network Layer. In SIGCOMM 98 Symposium Proceedings on Communications Architectures and Protocols, pages 303--313, 1988.

No context found.

K. Ramakrishnan and R. Jain. A binary feedback scheme for congestion avoidance in computer networks. ACM Transactions on Computer Systems, 8(2):158--181, 1990.

No context found.

K. K. Ramakrishnan and R. Jain. A Binary Feedback Scheme for Congestion Avoidance in Computer Networks. ACM Transactions on Computer Systems, 8(2):158--181, 1990.

No context found.

K.K. Ramakrishnan and R. Jain. A Binary Feedback Scheme for Congestion Avoidance in Computer Networks. ACM Transactions on Computer Systems, 8(2):158--181, 1990.

First 50 documents  Next 50

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC