
|
 |
| IBM Research
|
 |
Communications & Networking

|
|
|
"A
Hidden Semi-Markov Model for Web Workload Self-Similarity," Shun-Zheng
Yu, Zhen Liu,
Mark S. Squillante,
Cathy Honghui Xia, and Li Zhang, Proceedings of 21st IEEE International
Performance, Computing, and Communications Conference, p. 65-72,
2002.
Measurements
of Web workloads indicate that significant traffic variability is
present on a wide range of time scales. Characteristics such as
long-range dependence and self-similarity can have a significant
impact on the server performance. In this paper, we propose a hidden
semi-Markov model (HSMM) to characterize the user request patterns
of Web workloads. Hidden semi-Markov models have been well studied
and successfully applied to many engineering and scientific problems.
The advantage of using a HSMM is its efficient forward-backward
algorithms for estimating model parameters to best account for an
observed sequence. We show that our proposed HSMM model asymptotically
characterizes second order self-similar workloads when some duration
distributions of the hidden states are heavy-tailed. A recursive
formula is developed for estimating the Hurst parameter of self-similarity.
We validate our model and estimation methods with respect to two
sets of empirical data (requests per second) collected from two
different Web servers. We then use this model to generate self-similar
workloads that exhibit the same statistical properties. These measurements
show that we can use as few as 4 states together with a simple Poisson
process and heavy-tailed Pareto holding time distributions to accurately
model the Web workloads considered in this study.
|
|
"A
Label-switching Packet Forwarding Architecture for Multi-hop Wireless
LANs," Arup Acharya, Archan Misra, Sorav Bansal (Stanford University)
Channel
speeds for the IEEE 802.11 [2][4] family of standards continue to
increase: while the recently proposed 802.11a operates at 54 Mbps,
enhanced versions operating at speeds up to 108 Mbps are also under
investigation. Such high-speed LAN standards are expected to further
increase the popularity of wireless access to the backbone infrastructure
and eventually lead to the deployment of multi-hop, wireless networks,
where the wired backbone is reachable only via multiple wireless
hops. Potential examples of this include in-building wireless networks
in malls, hotels and apartment blocks, and community networks where
rooftop antennas are used to create an ad-hoc wireless access infrastructure
in specific residential communities.
A router in wired network typically requires multiple network interfaces
to act as a router or a forwarding node. In an ad-hoc multi-hop
wireless network on the other hand, any node with a wireless network
interface card can operate as a router or a forwarding node, since
it can receive a packet from a neighboring node, do a route lookup
based on the packet’s destination IP address, and then transmit
the packet to another neighboring node using the same wireless interface.
This paper investigates a combined medium access and next-hop address
lookup based on fixed length labels (instead of IP addresses), which
allows the entire packet forwarding operation to be executed within
the wireless NIC without the intervention of the host protocol stack.
Medium access schemes to date, such as IEEE 802.11, have been designed
implicitly for either receiving or transmitting a packet, but not
for a forwarding operation, i.e. receiving a packet from an upstream
node and then immediately transmitting the packet to a downstream
node as an atomic channel access operation. This paper proposes
a MAC protocol for packet forwarding in multi-hop wireless networks.
The proposed protocol builds on the IEEE 802.11 DCF MAC using RTS/CTS
and uses MPLS like labels in the control packets (RTS/CTS) to allow
the forwarding node to determine the next hop node while contending
for the channel. The throughput of this protocol is compared with
802.11 DCF MAC through simulation.
|
|
"An
Overview of the Bluetooth Wireless Technology," Chatschik
Bisdikian, IEEE Communications Magazine, Vol 39, No 12, pp 86-94,
December 2001.
This
paper won the Best Tutorial paper award given to an outstanding
tutorial paper published in any communications Society publication
in 2001. Dr. Bisdikian is an authority on Bluetooth wireless technology.
He is involved with the development of the Bluetooth specification
and serves as the vice-chair of the 802.15.1 task group that produced
the IEEE 802.15.1 standard for WPANs. The paper provides an excellent
introduction to the Bluetooth technology accessible to a wide audience
and provide references for a deeper study of the field.
|
|
"Inferring
Client Response Time at the Web Server," David P. Olshefski,
Jason Nieh, and Dakshi
Agrawal, Performance Evaluation Review, vol. 30, no. 1, p. 160-71,
2002, 06/2002.
As businesses
continue to grow their World Wide Web presence, it is becoming increasingly
vital for them to have quantitative measures of the client perceived
response times of their web services. We present Certes (CliEnt Response
Time Estimated by the Server), an online server-based mechanism for
web servers to measure client perceived response time, as if measured
at the client. Certes is based on a model of TCP that quantifies the
effect that connection drops have on perceived client response time,
by using three simple server-side measurements: connection drop rate,
connection accept rate and connection completion rate. The mechanism
does not require modifications to http servers or web pages, does
not rely on probing or third party sampling, and does not require
client-side modifications or scripting. Certes can be used to measure
response times for any web content, not just HTML. We have implemented
Certes and compared its response time measurements with those obtained
with detailed client instrumentation. Our results demonstrate that
Certes provides accurate server-based measurements of client response
times in HTTP 1.0/1.1 environments, even with rapidly changing workloads.
Certes runs online in constant time with very low overhead. It can
be used at web sites and server farms to verify compliance with service
level objectives.
|
|