Effect of physical layer bandwidth variation on TCP performance in cdma2000

Chaponniere, E.; Kandukuri, S.; Hamdy, W.

doi:10.1109/vetecs.2003.1207558

Cited by 8 publications

(8 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Conversely, for large windows (high TCP sending rates), it may be advisable to allocate high channel rates even at the expense of high bit error rates, since a low channel rate will inevitably result in packet loss due to congestion. On a related note, a previous study found that sharp bandwidth oscillations induced by rate adaptation of the RF scheduler in CDMA networks that are agnostic to TCP result in throughput degradation [3].…”

Section: Introductionmentioning

confidence: 93%

“…7 Since we do not account for time-outs in our model, we only simulated scenarios with large window size that result in few time-outs. Indeed, a previous study [3] has recommended the use of large window sizes in cellular networks to precisely avoid such time-outs due to bandwidth oscillations. The duration of each simulation run was 300 s. The bit error probability was held constant for the duration of the simulation, which is true under perfect power control.…”

Section: Adaptive Coding Evaluationmentioning

confidence: 99%

“…The set of data rates is obtained by choosing a combination of the RF control variables that produces the highest channel data rate under a particular channel condition 2 for a target FER. Similarly, scheduling policies typically assign the highest possible data rate allowed (for a given channel condition) from the set of given data rates that can clear the buffer backlog 3 (see [2] for a detailed characterization of this behavior). While ideal for inelastic constant rate applications, this resource allocation methodology can produce suboptimal performance of elastic applications and protocols, in particular TCP, that adapt their rate in response to feedback from the receiver.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

TCP-Aware Channel Allocation in CDMA Networks

Ghaderi

Sridharan

Zang

et al. 2009

IEEE Trans. on Mobile Comput.

View full text Add to dashboard Cite

Abstract-This paper explores the use of rate adaptation in cellular networks to maximize throughput of long-lived TCP sessions. Modern cellular networks incorporate RF technology that allows them to dynamically vary the wireless channel rate in response to user demands and channel conditions. However, the set of data rates as well as the scheduler's rate adaptation policy are typically chosen to maximize the throughput of inelastic connections. We focus on the problem of maximizing the throughput of TCP connections and propose a joint optimization of MAC and physical layer parameters with respect to TCP sending rate. In particular, we propose a simple TCP-aware channel scheduler that adapts the wireless channel rate to changes in the TCP sending rate and explore its performance for both single and multiple concurrent sessions. In the case of a single TCP session, we develop a fluid model of its steady-state behavior in such a system that adapts between two channel rates. The accuracy of the model, its utility in selecting optimal rates as well as the performance of systems with up to three channel rates are explored with ns-2 simulations. Our results indicate that a two-rate scheme improves TCP throughput by 15 to 20 percent over a system that does not exploit rate adaptation and that little additional benefit accrues from the addition of a third channel rate. Finally, we extend the framework to scenarios where bandwidth is shared by multiple TCP sessions. We propose two channel allocation algorithms, one relying on detailed TCP state information, the other not, and explore their performance through simulation. Our results indicate that TCP throughput is relatively insensitive to either channel allocation algorithm, and adaptive rate variation is the dominant factor in performance.

show abstract

Section: Introductionmentioning

confidence: 93%

Section: Adaptive Coding Evaluationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

TCP-Aware Channel Allocation in CDMA Networks

Ghaderi

Sridharan

Zang

et al. 2009

IEEE Trans. on Mobile Comput.

View full text Add to dashboard Cite

show abstract

“…User performance studies for CDMA2000 were published in many papers [4][5][6]. In [4], data service performance were evaluated for 3G-1X RTT system but alternative architecture or voice service were not addressed.…”

Section: Introductionmentioning

confidence: 99%

Performance modeling and evaluation of data/voice services in wireless networks

Kim

Lee

et al. 2006

Wireless Netw

View full text Add to dashboard Cite

Application-level performance is a key to the adoption and success of the CDMA 2000. To predict this performance in advance, a detailed end-to-end simulation model of a CDMA network is built to include application traffic characteristics, network architecture, network element details using the proposed simulation methodology. We assess the user-perceived application performance when a RAN and a CN adopt different transport architectures such as ATM and IP. To evaluate the user-perceived quality of voice service, we compare the end-to-end packet delay for different vocoder schemes such as G.711, G.726 (PCM), G.726 (ADPCM), and vocoder bypass scheme. By the simulation results, the vocoder bypass scenario shows 30% performance improvement over the others. We also compare the quality of voice service with and without DPS scheduling scheme. We know that DPS scheme keep the voice delay bound even if the service traffic is high. For data packet performance, HTTP v.1.1 shows better performance than that of HTTP v.1.0 due to the pipelining and TCP persistent connection. We may conclude that IP transport technology is better solution for higher FER environment since the packet overhead of IP is smaller than that of ATM for web browsing data traffic, while it shows opposite effect to the small size voice packet in RAN architecture. We show that the 3G-1X EV-DO system gives much better packet delay performance than 3G-1X RTT. The main conclusion is that end-to-end application-level performance is affected by various elements and layers of the network and thus it must be considered in all phases of the development process.

show abstract

“…The only experimental study we are aware of is the one presented in (Chaponniere et al, 2003) which evaluated the impact of bandwidth variation in CDMA2000 networks on the TCP timeout mechanism in a lab environment. The reader is referred to Chan and Ramjee, 2002) for relevant theoretical TCP models.Even less research exists on quantifying the extent each of the secondary factors influences the channel assigned rate.…”

Section: Problem Statementmentioning

confidence: 99%