R. Koodli and R. Ravikanth, "One-way loss pattern sample metrics." IETF RFC 3357, 2000.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....measurement packets [7] One way Packet Loss: The number of lost packets in a stream of packets from a source to a destination [2] One way Packet Loss Patterns: The patterns of packet losses in a stream of packets, e.g. the number of consecutive lost packets, the number of loss periods, etc. [14] Goodput Throughput: The rate perceived by a flow in bit s. In case of TCP flows, the goodput of the payload is measured. The document is structured as follows. The next section describes the general approach for QoS measurements in AQUILA and the derived measurement requirements. Section 3 ....

R. Koodli and R. Ravikanth. One-way loss pattern sample metrics. Internet Draft, IETF, November 2000. Work in progress.

....by few events, caused e.g. by link outages and route flappings. The Gilbert model thus only provides a limited insight with regard to the correlation of losses (loss burstiness) Several researchers provided additional intra flow loss models and metrics ( Par92, MFO98, OMF98, KR97, CT97, ZF96, KR00, NKT94, LNT96, KK98, LNT96] However most of these metrics are neither inter related nor well motivated. A law A law A D Analog input PCM D A Analog output silence detection delay adaptation Application Receiver Sender Transport Protocol Real Time G.729 G.729 Transport ....

.... Gamma m 1)o n 5 = k(m Gamma k 1)o k 2 o n A m = kmo k Gamma k(k Gamma 1)o k 2 k(k Gamma 1) o k m(m Gamma 1) nmo n Gamma m(m Gamma 1) p L Similar window based metrics were proposed also in ( OMF98, KR00, NKT94] The run length based model with a finite state space implies a geometric distribution for residing in the last state X = m. When we consider e.g. an estimation model order of m = 2 we can easily derive estimates for higher order model representations. As an example we consider the ....

[Article contains additional citation context not shown here]

R. Koodli and R. Ravikanth. "One-way Loss Pattern Sample Metrics ". Internet Draft, IETF IPPM Working Group, July 2000. ftp://ftp.ietf.org/internet-drafts/draft-ietf-ippm-loss-pattern-03.txt.

....burstiness. Therefore our goal is to establish feasible metrics that can suciently characterize the packet loss and delay processes and reliably predict perceived quality. We propose a joint use of the extended Gilbert model [19] 2 state being a special case) and inter loss distance (ILD) metric [12] to characterize temporal dependency in loss. The extended Gilbert model is suitable for describing loss run distributions, and the ILD metric is useful in capturing the burstiness between loss runs. To model temporal correlation of delay, we introduce a metric called conditional CDF. We analyzed ....

....Markov model assumes all past n events can a ect the future; whereas in an extended Gilbert model only the past (up to) n consecutive loss events will a ect the future. Therefore, it does not capture the burstiness or clustering between loss runs. However, we can use the inter loss distance metric [12] for this purpose. Figure 3 illustrates how the extended Gilbert model works. The system keeps a counter l, which is the number of consecutively lost packets, but it is reset whenever the next packet is delivered. The parameter to determine in an extended Gilbert model is P [X i jX i 1 to X i l ....

[Article contains additional citation context not shown here]

R. Koodli and R. Ravikanth. One-way loss pattern sample metrics. Internet Draft, Internet Engineering Task Force, June 1999. Work in progress.

....61, 62] but may on the other hand be less objectionable than spreading the same loss over a larger number of audio or video frames. Since audio and video packets are often closely spaced and small, loss correlation is more of an issue here than with many data applications [63, 64] IPPM defines [65] metrics for loss period lengths and for noticeable losses, i.e. losses separated by less than a threshold of uninterrupted packets. A similar metric is used by Maxemchuk and Lo [66] to evaluate packet audio performance. 3.5.3 Packet Reordering Under certain conditions, it is possible that ....

R. Koodli and R. Ravikanth, "One-way loss pattern sample metrics," Internet Draft, Internet Engineering Task Force, June 1999. Work in progress.

....has published several speci cations for measuring delay and loss. The following table is a partial summary: Document Reference Topics Covered RFC 2679 [2] one way delay measurement RFC 2680 [3] one way loss measurement RFC 2681 [4] two way (round trip) delay measurement ippm loss pattern [20] one way loss pattern characterization ippm ipdv [12] jitter measurement RFC 2679 has several key characteristics. First, it states that the clock di erence and drift at both sender and receiver must be synchronized or compensated for, but it does not specify how to achieve synchronization. ....

....(e.g. a 5 second timeout) The ippm ipdv draft [12] computes jitter by subtracting one way delays of two consecutive packets. It is equivalent to the RTP [40] de nition of jitter. This is also the de nition we use here for all the tables and gures in this paper. The ippm loss pattern draft [20] de nes several new metrics to capture packet loss patterns. These include loss distance (sequence di erence of current and previous lost packets) loss period (numbering of loss bursts and their lengths) noticeable loss rate (percentage of lost packets with loss distances smaller than a ....

R. Koodli and R. Ravikanth. One-way loss pattern sample metrics. Internet Draft, Internet Engineering Task Force, June 1999. Work in progress.

....Networks The IP Performance Metrics Working Group (IPPM) of IETF [IPPM] is conducting a major effort for the definition of a set of standard metrics that can be applied to evaluate data delivery services. IPPM considers QoS characteristics (such as delay or loss) independently [24] 1] 2] [16]. In spite of IPPM and other work, there isn t, as far as the authors know, a comprehensive way to measure QoS considering different characteristics altogether. The metric developed at LCT UC intends to address this challenge: to provide a broad and coherent view of the QoS given to applications, ....

....available for high priority traffic. The basic algorithm that supports the proposed scheduler is presented in Appendix A and detailed in [25] The dropper developed at LCT UC [26] includes also some important lessons learnt from experiments and from the work of other researchers [6] 18] 7] [16]. It avoids dropping TCP packets in order to protect TCP traffic from UDP traffic TCP has its own mechanisms for congestion avoidance triggered by packet drops whereas UDP has no such mechanisms. In addition, UDP and TCP packets are randomly dropped. Most important, the dropper removes memory ....

R. Koodli, R. Ravikanth, One-way Loss Pattern Sample Metrics, Internet Draft, October 1999. http://www.ietf.org/internet-drafts/draft-ietf-ippm-loss-pattern-02.txt

....scheduling disciplines. When executing drops, queue management systems should also consider the nature of the involved traffic. They should avoid dropping consecutive UDP packets belonging to the same flow, because the impact of loss will be dependent on the distance between dropped packets [9] for most applications that use this transport protocol. Consecutive dropping should also be avoided for TCP packets, in order to minimize the probability of eliminating packets belonging to the same TCP flow. Still related to the diverse nature of traffic, QM systems should deal with the growing ....

....the router prototype. The tolerance to packet bursts was evaluated measuring the number of packets dropped in sequence a higher tolerance to bursts means less packets dropped in sequence. The effectiveness of the dropper strategy was evaluated using a metric developed by Koodli and Ravikanth [9]. In general terms, the intention was to measure whether or not the prototype was able to spread out the drop of UDP packets using the referred loss distance stream metric. This metric determines a spread factor associated with the packet loss rate. Consider a stream of packets characterised by a ....

Koodli, R., Ravikanth, R.: One-way Loss Pattern Sample Metrics, Internet Draft, October 1999. http://www.ietf.org/internet-drafts/draft-ietf-ippm-loss-pattern-02.txt

....the packet was lost or not [1] Besides the packet generation time, we also record the instances when loss occurs at the common buffer. Loss distance is defined as difference in sequence numbers of two successively lost packets that may or may not be separated by successfully received packets [8]. Loss episode begins with a lost packet if the previous packet was successfully received. For example, if packets with sequence numbers 1, 4, and 6 are successfully received and packets 2, 3, and 5 are lost, then the first loss episode begins with packet 2 and ends with packet 3, while the second ....

R. Koodli and R. Ravikanth, "One-way loss pattern sample metrics," Internet draft, IETF, Oct. 1999, ftp://ftp.ietf.org/internet-drafts/draft-ietfippm -loss-pattern-02.txt .

....has published several speci cations for measuring delay and loss. The following table is a partial summary: Document Reference Topics Covered RFC 2679 [2] one way delay measurement RFC 2680 [3] one way loss measurement RFC 2681 [4] two way (round trip) delay measurement ippm loss pattern [20] one way loss pattern characterization ippm ipdv [12] jitter measurement RFC 2679 has several key characteristics. First, it states that the clock di erence and drift at both sender and receiver must be synchronized or compensated for, but it does not specify howtoachieve synchronization. ....

....(e.g. a 5 second timeout) The ippm ipdv draft [12] computes jitter by subtracting one way delays of two consecutive packets. It is equivalent to the RTP [40] de nition of jitter. This is also the de nition we use here for all the tables and gures in this paper. The ippm loss pattern draft [20] de nes several new metrics to capture packet loss patterns. These include #### ######## (sequence di erence of current and previous lost packets) #### ###### (numbering of loss bursts and their lengths) ########## #### #### (percentage of lost packets with loss distances smaller than a ....

R. Koodli and R. Ravikanth. One-way loss pattern sample metrics. Internet Draft, Internet Engineering Task Force, June 1999. Work in progress.

No context found.

R. Koodli and R. Ravikanth, "One-way loss pattern sample metrics." IETF RFC 3357, 2000.

No context found.

R. Koodli and R. Ravikanth, "One-way loss pattern sample metrics," Task Force, IP Performance Metrics Working Group, Internet Draft, Internet Eng., Nov. 2000.

No context found.

Koodli, R. and R. Ravikanth: 2000, `One-way loss pattern sample metrics'. Internetdraft, Internet Engineering Task Force. Expiration Date: December, 2000.

No context found.

Rajeev Koodli and Rayadurgam Ravikanth. OneWay Loss Pattern Sample Metrics, August 2002. RFC 3357.

No context found.

R. Koodli and R. Ravikanth, #One-way loss pattern sample metrics," Internet Draft, IETF IPPM Working Group, June 1999. ftp:##ftp.ietf.org#internet-drafts#draft-ietf-ippm-loss-pattern-01.txt.

No context found.

R. Koodli and R. Ravikanth, "One-way loss pattern sample metrics," Internet Draft, IETF IPPM Working Group, June 1999. ftp://ftp.ietf.org/internet-drafts/draft-ietf-ippm-loss-pattern-01.txt.

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