ClubNet Spring 2005/6
23/03/06
14:30, 1003
|
Dr.
Philip M. Merlin Memorial Lecture and Prize Award
Broadband Wireless:
Promise and Reality
Prof. Andrew J. Viterbi
Viterbi Group, LLC and Distinguished Visiting Professor, Technion
*The Merlin Prize will be awarded to a distinguished student; refreshments
will be served at 14:00
|
Organized by Evgeny Bolotin and
Ron Banner.
The End of the
Tier-Based Architecture
Nati Shalom
CTO, GigaSpaces
The emergence of powerful and new commodity hardware and the
introduction of SOA/grid architectures tout the promise of achieving true
linearly-scalable systems at a lower cost. Unfortunately, these new systems and
architectures are not aligned with the existing tier-based approach, which is by
definition centralized and static. In this paper we will introduce a new
approach - Space-Based Architecture (SBA) for transforming existing tier-based
applications into linearly and dynamically scalable services. SBA represents a
new model that combines distributed caching ("Data Grid"), content-based
messaging ("Messaging Grid") and parallel processing ("Processing Grid"). These
new middleware components are implemented on top of a distributed shared memory
space that provides common clustering, high availability, location transparency
and consistency services across all tiers.Unlike most traditional approaches,
with SBA scalability is handled implicitly within the middleware layer, enabling
applications to Write Once their business logic and effectively Scale Anywhere
They can both scale up within a machine and scale out across machines
transparently and dynamically. The SBA design is optimized for high-performance
stateful SOA applications.
Parallel and Distributed
Systems: Theoretical and System Perspectives
Liran Liss
EE Department, Technion
We revisit key performance issues and assumptions on which modern
parallel and distributed systems are constructed, while taking both theoretical
and practical points of view. Specifically, we focus on two representative forms
of computation: tightly-coupled clusters for parallel computing and extremely
large-scale distributed systems. In tightly-coupled clusters,
the communication bottleneck at the end-nodes themselves is often the major
cause for poor performance. While the introduction of System Area Networks (SANs)
alleviated this bottleneck, applications seldom realize raw hardware
capabilities, mainly due to substantial mismatches between the software
architecture and hardware. One source for such mismatches is
the strong semantics of standard messaging APIs, which are not always required
by the application. Therefore, many common overheads can be tackled upfront by
deviating from these APIs. Using our full-fledged prototype of a demanding
Distributed Shared Memory (DSM) system, we show how network and OS primitives
can be integrated in the kernel to provide a high-performance platform that is
well matched to application semantics. An alternative approach that we explore
entails avoiding most overheads using theoretical memory models that are better
matched to the hardware's semantics. Specifically, we assess implementation
tradeoffs of the Bulk Synchronous Parallel (BSP) model, which endorses
coarse-grain synchronization. In extremely large-scale
systems, in contrast, the varying communication latencies between nodes and
dynamic changes in topology play a far more important role, and the
ever-increasing demand for scalability has driven locality considerations into
every aspect of algorithm design. Unfortunately, there are many important
problems that cannot be solved in a local manner per se, i.e., O(1) complexity
in problem size for all instances. In addition, the common measures for
characterizing the potential efficiency of algorithms and that actually achieved
by them, namely worst- and average-case (over problem instances) complexities,
fail to capture the essence of real problems in a meaningful way.
Concentrating on global aggregation problems, which constitute an important
application in many large-scale systems, we define a new metric on problem
instances, Veracity Radius (VR), which captures the inherent possibility to
compute them locally. We prove that VR yields a tight lower bound on both
output-stabilization and quiescence times, and provide an efficient algorithm
whose performance is proportional to VR for every problem instance rather than
the graph size. Finally, we demonstrate the benefits of our instance-local
approach for highly dynamic networks with two practical algorithms: an algorithm
for majority voting and a scalable quota management protocol for Grid computing
environments.
*
סטודנט לתואר שלישי בהנחיית דר' יצחק בירק ופרופ' אסף שוסטר
.
Broadband
Wireless: Promise and Reality
Prof. Andrew J. Viterbi
Viterbi Group, LLC and Distinguished Visiting Professor, Technion
Dr. Philip M. Merlin Memorial Lecture and Prize Award
The Merlin Prize will be awarded to a distinguished student; refreshments will
be served at 14:00
Motorola
Seamless Mobility
Saul Emile
Global Sales Manager, Motorola Israel Seamless Mobility allows
subscribers to move between environments, using multiple mode devices, while
accessing a common set of applications and services available wherever and
whenever they want. It is a revolution that will affect everything from the
design, integration and deployment of applications, to the evolution of network
infrastructure and operations, to the devices used to access those networks.
Devices will be able to move between heterogeneous networks. For example a dual
mobile device that can be used on both Wi-Fi and cellular networks. That means
the subscriber's home phone can be the same as the mobile phone or, in the
office a, a desk phone could transfer a call to an in building Wi-Fi LAN network
when the subscriber leaves the office. Content will be able to move seamlessly
across devices. Achieving this vision will transform industries and enhance
convenience, productivity, fun and safety.
The new challenge
of IPTV with VOD, Web Browser, VoIP and video telephony everywhere
Danny On
Co-Founder & CEO, U Control Me Ltd Innovative solutions for Video &
Voice through Broadband & 3G Cellular (IP STB, VOD, VoIP, 3G) .The new challenge
of IPTV with VOD, Web Browser, VoIP and video telephony every where.
- WHAT are the applications that the market wants?
- WHEN will it happen?
- WHAT are the standards?
- WHICH standards will dominate?
- WHO are the players?
- WHAT is the new architecture?
- Today’s demand
- What is VoIP & Video Telephony ?
- A case study
Lecture slides [PDF]
Random Sampling
from a Search Engine's Index
Maxim Gurevich
EE
Department, Technion
We revisit a problem introduced by Bharat and Broder almost a decade ago: how to
sample random pages from a search engine's index using only the search engine's
public interface? Such a primitive is particularly useful in creating objective
benchmarks for search engines. The technique of Bharat and Broder suffers from
two well recorded biases: it favors long documents and highly ranked documents.
We introduce two novel sampling techniques: a lexicon-based technique and a
random walk technique. Our methods produce biased sample documents, but each
sample is accompanied by a corresponding “weight”, which represents the
probability of this document to be selected in the sample. The samples, in
conjunction with the weights, are then used to simulate near-uniform samples. To
this end, we resort to four well known Monte Carlo simulation methods: rejection
sampling, importance sampling, the Metropolis-Hastings algorithm, and the
maximum-degree method. We analyze our methods rigorously and prove that under
plausible assumptions, our techniques are guaranteed to produce near-uniform
samples from the search engine's index. Experiments on a corpus of 2.4 million
documents substantiate our analytical findings and show that our algorithms do
not have significant bias towards long or highly ranked documents. We use our
algorithms to collect fresh interesting data about the relative sizes of MSN
Search, Google and Yahoo!. Joint work with Ziv Bar-Yossef.
Enterprise Data
Center Evolution
Michael Kagan
VP
of Architecture, Mellanox Technologies The growing
demand for computing and storage services in the Enterprise Data Center results
in exponential growth of service delivery cost. The emerging trend to address
this cost issue is resources' virtualization and development of mechanisms to
deliver resources per demand. The talk will overview current solutions for
servers' and IO virtualization and discuss emerging solutions for the Virtual
Enterprise Data Center.
Packet-Mode
Emulation of Output-Queued Switches
David Hay
CS
Department, Technion
Most common network protocols (e.g., the Internet Protocol) work with variable
size packets, whereas contemporary switches still operate with fixed size cells,
which are easier to transmit and buffer. This necessitates packet segmentation
and reassembly modules, resulting in significant computation and communication
overhead that might be too costly as switches become faster and bigger.
In this talk we consider an alternative mode of scheduling,
in which packets are scheduled contiguously over the switch fabric.
Specifically, we study such packet-mode scheduling for the combined input output
queued (CIOQ) switch architecture and investigates its cost.
We introduce frame-based schedulers that allow a packet-mode
CIOQ switch with small speedup to mimic an ideal output-queued switch with
bounded relative queuing delay. The schedulers are pipelined and are based on
matrix decompositions.
Joint work with Isaac Keslassy (EE) and Hagit Attiya (CS).
Two Talks For
The Price of One: A Survey of Network Coding Algorithms, and Talking About a New
One
Sidharth Jaggi
MIT
Network coding, a powerful paradigm for the design of codes for information
transmission over networks (explored in an excellent paper by Ahlswede et al in
2000) has attracted considerable attention over the last few years. The idea is
simple -- all nodes in a network (rather than just the sources and sinks) can
and should perform non-trivial operations (rather than just copying and
forwarding). The combinatorial explosion in interest in the topic is due in part
to the elegant mathematics that powers network coding algorithms, but also due
to the fact that many information theoretic problems for complex networks can be
shown to have rate-regions that are simple to describe, and algorithms that are
corresponding low in design and implementation complexity.
This talk is divided into two parts.
The first half presents background concepts and a short survey of algorithms
from the field of network coding.
In the second part of the talk we focus on the design and analysis of
error-correcting codes for networks. We phrase the problem in a worst-case,
adversarial model; in particular, we design codes to transmit information over a
network, some subset of which is controlled by a malicious adversary. In
addition, the topology of the network is unknown to the network nodes prior to
communication, and the network nodes and links of which are prone to failures.
The computationally unbounded, hidden adversary knows the message at the source,
knows the network topology and codes being used, and can observe and change
information over the part of the network he controls. The network nodes do not
share resources such as shared randomness or a private key. We describe several
algorithms (tailored for different flavours of the problem) that can achieve a
rate-region that can be shown to be information-theoretically optimal.
A Legend-based
Proof of the GHS Minimum-spanning Tree Algorithm
Prof.
Yoram Moses
EE
Department, Technion This talk will describe a
new proof of correctness for the celebrated Minimum Spanning Tree protocol of
Gallager, Humblet and Spira [1983].
Both the protocol and the quest for a natural correctness proof have had
considerable impact on the literature concerning both network protocols and
verification. We present an invariance proof that is based on a new
intermediate-level abstraction of the protocol. One novel aspect is that
intermediate-level configurations serve as a legend describing the state of the
GHS computation. This provides a powerful tool for both the statement and the
proof of properties of the algorithm. The result is the first proof that follows
the spirit of the informal arguments made in the original paper.
This is joint work with Benny Shimony.
Multi-Path
Routing
Ron Banner
EE Department, Technion Unlike traditional routing
schemes that route all traffic along a single path, multipath routing strategies
split the traffic among several different paths. In spite of some well known
benefits of multipath routing, its properties in theory and in practice got
relatively little attention from the research community. In this study, we
investigate several fundamental aspects in the context of multipath routing. We
establish a number of positive and negative theoretical results regarding the
capabilities of multipath routing that give rise to its efficient deployment in
practical settings.
Our study investigates multipath routing from several
perspectives. One is the purpose of employing multipath schemes. We primarily
focused on the major goals of congestion avoidance and enhanced survivability,
from an algorithmic point of view. In addition, we considered other algorithmic
issues of multipath routing, such as the need for topology control in wireless
networks when multipath solutions are employed. Another aspect is the available
information on the problem's input. While we mainly focused on offline multipath
problems, we also considered their online versions, where there is no a-priori
knowledge regarding future demands. A third aspect that is considered in this
study is the cooperation (or lack thereof) among decision makers. While we
mainly focused on global optimization, we also explored the fundamentally
important case where multipath routing is performed by selfish decision makers.
Finally, while most of this study provides evidences to the major potential of
multipath routing, we have characterized several important cases where multipath
routing schemes offer only limited benefits and should therefore not be favored
over the traditional, single-path routing approach.
Under the supervision of Prof. Ariel Orda
IMS - IP
Multimedia Subsystem
Ze'ev Likwornik
Mgr Business Development, Cisco IMS - IP Multimedia
Subsystem - is a standardized Next Generation Networking architecture for
telecom operators that want to provide mobile and fixed multimedia services. It
uses a Voice-over-IP (VoIP) implementation based on a 3GPP standardized
implementation of SIP, and runs over the standard Internet Protocol (IP). Zeev
will present past, current and future of IMS with some emphasize on Marketing
and Business aspects.
|
Login
