-
[show abstract]
[hide abstract]
ABSTRACT: In networks carrying large volume of traffic, accurate traffic characterization is necessary for understanding the dynamics and patterns of network resource usage. Previous approaches to flow characterization are based on random sampling of the packets (e.g., Cisco's NetFlow) or inferring characteristics solely based on long lived flows (LLFs) or on lossy data structures (e.g., bloom filters, hash tables). However, none of these approaches takes into account the heavy-tailed nature of the Internet traffic and separates the estimation algorithm from the flow measurement architecture.In this paper, we propose an alternate approach to traffic characterization by closely linking the flow measurement architecture with the estimation algorithm. Our measurement framework stores complete information related to short lived flows (SLFs) while collecting partial information related to LLFs. For real-time separation of LLFs and SLFs, we propose a novel algorithm based on typical sequences from information theory. The distribution (pdf) and sample space of the underlying traffic is estimated using the non-parametric Parzen window technique and likelihood function defined over the Coupon collector problem. We validate the accuracy and performance of our estimation technique using traffic traces from the internal LAN in our laboratory and from National Library for Applied Network Research (NLANR).
IEEE Transactions on Parallel and Distributed Systems 02/2009; · 1.40 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: Today's distributed computing systems incorporate different types of nodes with varied bandwidth constraints which should be considered while designing cost-optimal job allocation schemes for better system performance. In this paper, we propose a fair pricing strategy for job allocation in bandwidth-constrained distributed systems. The strategy formulates an incomplete information, alternating-offers bargaining game on two variables, such as price per unit resource and percentage of bandwidth allocated, for both single and multiclass jobs at each node. We present a cost-optimal job allocation scheme for single-class jobs that involve communication delay and, hence, the link bandwidth. For fast and adaptive allocation of multiclass jobs, we describe three efficient heuristics and compare them under different network scenarios. The results show that the proposed algorithms are comparable to existing job allocation schemes in terms of the expected system response time over all jobs
IEEE Transactions on Parallel and Distributed Systems 04/2007; 18(3):289-306. · 1.40 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: This paper proposes a new measurement architecture and associated traffic estimation algorithm called FastFlow that uses the heavy-tailed nature of Internet traffic in order to distinguish packets belonging to short lived flows (SLFs) and long lived flows (LLFs). While complete information is stored for SLFs, only partial information related to LLFs is collected using systematic sampling. The absence of data points in LLFs is approximated using a likelihood function defined over the coupon collector problem and the distribution of underlying traffic estimated using the non-parametric Parzen window technique. We validate the performance of our approach using traffic traces collected from our lab and observe that the estimated statistics match the observed traces with high accuracy.
Distributed Computing Systems, 2006. ICDCS 2006. 26th IEEE International Conference on; 02/2006
-
[show abstract]
[hide abstract]
ABSTRACT: In this paper we study the causal behavior of Rayleigh fading wireless channel and a single queue system buffer using the finite state Markov chain (FSMC) and the finite buffer fluid flow model. In the process, we propose a state partitioning scheme for capturing the wireless channel realities based on the combined effect of level crossing rate (LCR), average fade duration (AFD) and the average value of the signal SNR observed at the output of the matched filter or correlation demodulator. Relevant system parameters captured using such a scheme is linked with the fluid flow traffic model to create a modulated Markov process (MMP) that is used to calculate the packet loss rate and effective bandwidth of the system.
Modeling and Optimization in Mobile, Ad Hoc, and Wireless Networks, 2005. WIOPT 2005. Third International Symposium on; 05/2005
-
[show abstract]
[hide abstract]
ABSTRACT: In recent years, wide bandwidth and low cost wireless LAN (WLAN) technology has emerged as a competitive choice for high speed wireless Internet access. To support the Internet mobility for the frequently moving mobile nodes in the local WLAN environment, Mobile IP has been found to be inefficient due to its high location update and packet delivery cost. This paper proposes an alternative architecture for network-layer mobility support using dynamic client configuration and transparent proxy mechanism. The attractive feature of our solution is that it requires no change in the legacy network infrastructure. We have implemented a prototype of the proposed architecture, and have taken experimental measurements to evaluate its performance. Experimental results show that the proposed architecture is capable of supporting IP mobility with the handoff delay less than a second. In addition, we have done an analytical study to test the scalability of the proposed architecture. Results show that performance of the architecture is not severely affected by the user mobility and a commonplace router is capable of supporting a considerable number of such users.
Network Computing and Applications, 2003. NCA 2003. Second IEEE International Symposium on; 05/2003
