ClubNet - Spring 2003

Organized by Omer Gurewitz and Gil Zussman.

Clubnet Main Page
Clubnet Winter 2002/3 lectures
ClubNet 2001/2 lectures
ClubNet 1999 lectures
ClubNet 1997/8 lectures

The IP Telephony Signaling Maze
Dr. Yoni Malachi
CTO and VP R&D, RADVISION

In less than a decade Voice over IP technology and products have been implemented both within enterprises and as a backbone network for new voice carriers.

In this introductory talk we will review the basics of the VoIP signaling protocols and compare them. The presentation will discuss H.323, SIP, Megaco, and MGCP comparing their main features and the supported architectures.

No prior knowledge of Voice over IP will be assumed.

Optimal Content Location in IP Multicast Based Overlay Networks
Oren Unger
Staff CAD Engineer, Zoran Microelectronics Ltd.

An overlay network is a virtual network topology constructed on top of the Internet structure. The architecture of overlay networks should encourage high-performance, high-scalability and reduced costs. This becomes more crucial when communication, storage costs as well as service latencies grow with the exploding amounts of data exchanged and with the size and span of the overlay network. For that end, multicast methodologies can be used to deliver content from regional servers to end users, as well as for the timely and economical synchronization of content among the distributed servers. Another important architectural problem is the efficient allocation of objects to servers to minimize storage, delivery and update costs.

In this work, we suggest an IP multicast based architecture and address the optimal allocation and replication of objects that are both consumed and updated. Our model network includes consumers which are served using IP multicast and/or IP unicast transport and media sources which update the objects within the servers using multicast. General costs are associated
with distribution (download) and update traffic as well as the storage of objects in the servers.

We address the general object allocation problem on tree networks and present optimal algorithms for subsets of the problem. We also suggest a heuristic solution for a general graph network, for the multicast distribution and multicast update problem.

Joint work with Prof. Israel Cidon.

Presentation

Dr. Philip M. Merlin Memorial Lecture
Is IP the Ultimate Future Network?
Zohar Zisapel
Chairman of RAD Data Communications, Ltd.

The internet network and the IP protocol had promised a lot (nearly everything) during the late nineties. Now several years later the bubble had burst and we all who had been so sanguine before are now feeling malaise with our feet deep in quagmire. Are we more realistic now ? Did the Internet cheat us all ? Does IP have a big past ahead of it ? if so where are communications networks heading ? The lecture will try to answer those questions. I will be very happy if participants will express their view on this subject and transform the lecture to an open discussion.

Presentation
Video (available only to Technion users)

On Multicast Trees: Structure and Size Estimation
Osnat Mokryn
School of Engineering and Computer Science, The Hebrew University of Jerusalem

This work presents a thorough investigation of the structure of multicast trees cut from the Internet and power-law topologies. Based on both generated topologies and real Internet data, we characterize the structure of such trees and show that they obey the rank-degree power law; that most high degree tree nodes are concentrated in a low diameter neighborhood; and that the sub-tree size also obeys a power law.

Our most surprising empirical finding suggests that there is a linear ratio between the number of high-degree network nodes, namely nodes whose tree degree is higher than some constant, and the number of leaf nodes in the multicast tree (clients). We also derive this ratio analytically, and show that it holds for trees cut from the Internet. Based on this finding, we develop the Fast Algorithm, that estimates the number of clients, and show that it converges faster than one round trip delay from the root to a randomly selected client.

Joint work with Danny Dolev and Yuval Shavitt

Presentation

Crossing IP Telephony Barriers
Israel Gal
President and CEO, BOScom

For many years, up until the year 2000, IP Telephony was hailed as the immediate successor to traditional telephony networks.
Like many other new technologies, IP Telephony did not take the market by storm.
This session will discuss the barriers that have prevented rapid adoption of IP telephony in the market, and the solutions and strategies required to overcome these barriers.

Presentation

The Cellular Industry - Forward Looking Comments
Ayal Bar-David
General Manager of QUALCOMM Israel

We are uncertain about the engineering direction of the cellular industry, yet, in this presentation, we will try to explain what the driving forces are, and how the leaders in the industry are trying to calculate their steps. We will start off with a short overview of the current status of the industry, from that stand point of technology, standards and resources, and try to understand the motivation for the industry to move forward.  We will then present some technological challenges that are currently facing the industry, and show some concepts on how the leaders are moving forward in the area of Antennas and Propagation and Networking issues, such as Quality of Service.
 

Deterministic Distributed Resource Discovery
Prof. Shay Kutten
Dept. of Industrial Engineering and Management, Technion

The resource discovery problem was introduced by Harchol-Balter, Leighton and Lewin. They developed a number of algorithms for the problem in the weakly connected directed graph model. This model is a directed logical graph, that represents the vertices' knowledge about the topology of the underlying communication network.

We describe a deterministic algorithm for the problem in the same model, with improved time, message, and communication complexities. Each previous algorithm had a complexity that was higher at least in one of the measures. Specifically, previous deterministic solutions required either time linear in the diameter of the initial network, or communication complexity $O(n^3)$ (with message complexity $O(n^2)$), or message complexity $O(|E_0| \log n)$ (where $E_0$ is the arc set of the initial graph $G_0$). Compared to the main randomized algorithm of Harchol-Balter, Leighton, and Lewin, the time complexity is reduced from $O(\log^2n)$ to $O(\log n )$, the message complexity from $O(n \log^2 n)$ to $O(n \log n )$, and the communication complexity from $O(n^2 \log^3 n)$ to $O(|E_0|\log ^2 n )$.

Our work significantly extends the connectivity algorithm of Shiloach and Vishkin which was originally given for a parallel model of computation. Our result also confirms a conjecture of Harchol-Balter, Leighton, and Lewin, and addresses an open question due to R. Lipton.

Finally, we show how to translate our protocol to asynchronous networks while retaining similar complexities.

Presentation

Scalable Methods for Provisioning and Restoration of QoS Paths and Trees
Alex Sprintson
Dept. of Electrical Engineering, Technion

Communication networks are expected to meet a growing volume of requirements and support new applications such as multimedia streaming and video conferencing. Two essential requirements are support of Quality of Service (QoS) and resilience to failures. Supporting QoS connections requires the existence of  a routing mechanism that identifies a suitable routing topology, that is, a path for unicast and a tree for multicast. Resilience to failures, on the other hand, is achieved by  provisioning a restoration topology, that is, a set of restoration paths (or trees), each protecting a component of the primary topology. Upon failure of a network element in the primary path (tree), traffic is immediately switched to a restoration path (tree).  The above objectives, coupled with the need to minimize the global use of network resources, imply that the cost of both the primary path and the restoration topology should be a major consideration of the routing process.

In this talk we present an overview of our research on scalable methods for provisioning and restoration of QoS paths and trees. We then focus on the fundamental problem of identifying trees that satisfy additive QoS constraints (such as delay) at minimum cost. This problem has been extensively studied. However, existing solutions have either relied on heuristic approaches or considered special cases, such as the case where the delay and cost of each link are identical. We propose a novel algorithmic scheme, with proven performance guarantees, for this fundamental multicast problem. Effectively, this scheme allows to obtain an approximate solution to this problem out of any given approximate scheme of its (simpler) unconstrained directed version, with about identical (epsilon-close) performance guarantees.

Presentation

High-End IP Routers = Modern Supercomputers
Dr. Daniel E. Lenoski
VP of Engineering, Routing Technology Group, Cisco Systems

Growth of internet traffic together with advances in fiber optic bandwidth has created huge demand and supply of wide-area network bandwidth.  High-end internet routers are key to intelligently switching this exponentially increasing bandwidth.  This
has forced radical changes in router technology over the last 10 years.  This talk will highlight the key technology required for high-end routers and compare with equivalent usage in high-end supercomputing.   We will show that by virtually all measures of system capacity and sophistication, today's high-end routers are truely supercomputers.

Presentation
Video (available only to Technion users)

Bluetooth and Bluetooth Networking
Nathan Bruker
Senior Director, Technology and Strategy, IXI Mobile, Ltd.

Bluetooth technology provides the physical and logical layers for short distance wireless communications. Through an intensive use of standardized "profiles", the Bluetooth special interest group (Bluetooth SIG) tries to address different usage model like cable replacement, dial-up, LAN profile and object push (OBEX). The ultimate goal is to enable true networking between different devices composing a personal network around the user (Personal Area Network or PAN).

As new devices are coming to the market, this vision encounters practical limitations such as system concept and design, interoperability, and standardization.

The presentation will give a high level description of the Bluetooth standard, describe today limitations and than provide an overview of IXI Mobile's commercial solution that tries to solve some of those problems.

Presentation
Video (available only to Technion users)

Technion Computer Network - Today and Tomorrow
Dr. Dani Arbel
Network Manager, Taub Computer Center, Technion

The existing infrastructure of the technion comp. network will be described. The design goals and considerations will be presented, with a discussion about near future improvements.