Research Interests

Broadband packet networks including IP and ATM, accelerated simulation, queueing theory, packet level measurements, traffic control and network performance.

Current and recent PhD students

Ling Xu - Planning Simulation Run Length in Packet Queues in Communications Networks. 2013

Syeda Samana Naqvi - Packet level measurements over wireless. 2011

Vindya Amaradasa - Traffic aggregation techniques for non-FIFO schedulers. 2008

Maheen Hasib: Analysis of packet loss probing in packet networks . June 2006

Shaowen LU: A systematic multi-level abstraction approach to error constrained time-stepped accelerated simulation for MANETs . 2006

Sharifah Ariffin: Accelerated simulation of a packet buffer with a non-FIFO scheduler . March 2006

Ali Raza: Layered Space Time Architectures For Mimo Wireless Channels . 2006

Chi Ming Leung: Non-Intrusive Measurement in Packet Networks and its Applications . February 2004

Ho I (Athen) Ma: Accelerated Simulation of Power-Law Traffic in Packet Networks . September 2003

Robert Stewart, End-toEnd Delay Analysis for Small/Medium Scale IP Networks November 2002

Arif Al-Hammadi Intelligent Techniques for VBR Traffic Control in ATM Networks . March 2000

Tijana Timotijevic System Level Performance of ATM Transmission over a DS-CDMA Satellite Link April 1999

Current and recent projects

Optimal design of performance measurement experiments for complex, large-scale networks (DOENET)

Principal Investigator: Dr J A Schormans

Other Investigators: Professor SG Gilmour Professor JM Pitts

Partner Project: EP/G013993/1

Title: Optimal design of performance measurement experiments for complex, large-scale networks

Principal Investigator: Dr AW Moore (Computer Lab, Cambridge University)

Network measurement may have social, engineering or commercial motivations. Typical social motivations include the need to understand popular usage and to answer questions about social-networking sites; other possibilities include use as part of legal evidence to illustrate the (lack of) impact of regulation upon the popularity of peer-2-peer networks. The engineering application of measurement is an integral part of the network optimisation process / both to provide a baseline of pre-optimised performance and to quantify improvements. Another area of application is as part of measurement-based algorithms for controlling network access. Commercially, measurement is critical to guaranteeing Service Level Agreements (SLAs) between network and service providers and customers. These SLAs provide enforceable guarantees on the upper bounds of packet level performance; e.g. they state that mean delay, delay variation (sometimes called "jitter") and packet loss probability will not exceed a specified value when measured over an agreed period. However a critical problem is that network traffic and topologies are highly variable and this tends to create difficulties in measuring accurately. For this reason measurement can be prone to very large errors in estimating end-to-end delay (both mean delay and jitter) and packet loss rates. Indeed the optimal measurement of packet level performance is a challenging open problem in engineering mathematics. Current measurement methods are not actually designed to provide the maximum information from the minimum data set. In this project the crucial step is to view all network measurements as numerical experiments, in which random processes are sampled, and the sampling is constrained by the resources available, e.g. bandwidth. In this way we are then able to apply the Statistical Design of Experiments (DOE) to network measurement experiments. DOE techniques have been very successfully applied in linear and static environments, mainly in biological and some industrial contexts. Most work on DOE has assumed static processes, or deals only with the static aspects of the processes, but network traffic and topologies are highly variable and nonlinear. The first work on DOE for models which are solutions of nonlinear differential equations was in the field of chemical kinetics, co-authored by a member of the Statistics Group at Queen Mary. Subsequent work at Queen Mary has further developed DOE for nonlinear models, or nonlinear functions of the parameters in a linear model. Although the models involved are fairly small-scale compared with those arising in networks, having a single input variable and no feedback, this has provided a good starting point. In parallel with the DOE thread at Queen Mary (Statistics Group and Networks Group), the University of Cambridge (Computer Lab) builds on experience in techniques for packet (or packet flow) classification. Such techniques for the classification of network traffic have previously used features derived from streams of packets. Such feature collections are often huge (200+), and can range in complexity from Fourier-Transformations and quartile statistics to mean and variance of packet inter-arrival times and the number of TCP SACK packets. Classification accuracy is often good, but with the disadvantage of complexity and cost. In this project such previous experience with lightweight application classification schemes are re-oriented towards learning the traffic characteristics that are critical in their influence on delay and loss performance. This approach to focus on actual experimental observations; a better approach than relying on simple queue models that have inbuilt (and limiting) distributions chosen mainly to allow the resulting system of equations to be solved. The combination of DOE and machine learning promises a real step towards solving the problem of the optimal measurement of packet level performance.

 

EPSRC Reference: EP/F033133/1

Title: Building a New Community: Modelling, Visualisation and Verification of Large Scale Systems

Principal Investigator: Dr J Schormans

Other Investigators: Dr H Grossmann

Dr P Oliva

Dr JE Cater

Professor PW McOwan

Professor JM Pitts

Department: Electronic Engineering

Organisation: Queen Mary, University of London

 

Abstract: At Queen Mary University of London the team of investigators are all engaged in researching solutions to many of the problems that arise from large scale systems. These include: - Network engineering for the global-scale communications networks that support the Internet. The sheer size and complexity of the Internet's underlying networks create a host of new problems for engineers and computer scientists. Examples include the correct use of network simulators and the accurate me as urement of end-to-end performance across global-scale networks in order to provide customer guarantees. - Fluid mechanics of large-scale physical systems, e.g. aircraft bodies. Engineers study the generation of sound created by turbulent flow adjacent to a solid surface, which create some of the interior noise experienced by p as sengers in moving vehicles, particularly in aircraft and submarines. Experimental me as urements of these flows typically produce thousands of data streams requiring intensive processing to reveal the underlying modes of activity and the evolution of flow patterns. - Software systems with millions of lines of code, e.g. the operating systems of Windows, Linux or Internet Explorer. Most people have experienced the frustration of finding 'bugs' in their computer applications. As the number of lines of codes required to create these applications grows and grows these challenges get ever harder. - Next generation computers inspired by biological systems of immense size and complexity. The more you look at nature the more obvious it is that natural systems have already solved many of the problems that are faced by computer scientists. An example is the human eye-brain system, which can be viewed as a biological computer system of enormous size and complexity. In all of these research are as , because of the huge underlying system state-space and variability, researchers are faced with three key challenges: how to validate their network models, how to know that the results of any single system-wide experiment can be reproduced by other researchers and how to best visualise the systems under study. To create a truly interdisciplinary approach to these issues, engineers, electronic engineers and computer scientists are joined by researchers from the school of Mathematical Sciences , bringing expertise in the mathematics of large-scale nonlinear systems, and the optimal design of experiments. The purpose of this project is to facilitate new multidisciplinary interactions addressing these problems, by bringing together different perspectives through a range of project events. Our programme of events is aimed at overcoming discipline boundaries, allowing us, and our colleagues, to think through the possibilities, resulting in new multidisciplinary research project proposals to EPSREC. We intend to begin by b as ing our thinking on the fact that, while each research group uses techniques specific to their field, the underlying scientific questions remain largely the same (how to model, verify and visualise). Our approach to these challenges is to direct attention at specific are as , including: - Reproducibility of large-scale system experiments - User response to quality of large-scale engineered systems - Visualisation of system behaviour - Techniques for multi-scale model abstraction

Starts: 03 January 2008 Ends: 02 January 2011 Value (£): 247,978

Scheme: Standard Research

 

-----------------------------------------------------------------------------------------------------------------

EPSRC Reference: GR/S93714/03

Title: IPv4 and IPv6 Performance and QoS - 46PaQ

Principal Investigator: Dr AW Moore

Other Investigators: Dr J Schormans

Dr I Pratt

 

Department: Computer Laboratory

Organisation: University of Cambridge

 

Abstract: As demand for network capacity incre as es, the traditional practice of over-provisioning of the network becomes impractical. So, there is much ongoing work to define new protocols and mechanisms for high-speed, QoS-controlled networking within Internet Protocol (IP) b as ed environments. This includes work on QoS mechanisms and services, congestion control mechanisms and work on new transport protocols for specific purposes. However, as these new mechanisms have been developing, there h as not been much activity in trying to observe and analyse the behaviour of these systems working together, within a real, integrated networked environment. With the rapidly incre as ing deployment of IPv6, it is also vital to consider carefully the differences in behaviour in the use of these mechanisms compared to IPv4. In this project, we will have two broad are as of work. Firstly we will, through experimentation and analysis, examine how DIFFSERV, ECN and decentralised reservation can be made to operate together in a very high-speed IPv4 and IPv6 environment to support TCP- and UDP-b as ed applications. Secondly, we will propose ways in which such networks can be instrumented in order to provide performance and operational data to network operators as well as users and applications through appropriate APIs and using network monitoring equipment via configurable sampling techniques, which will also be developed in the project. We will examine the performance of the system in a real networked environment operating at very high speeds (several Gb/s). To stretch the network services, we intend to test with selected applications from the e-Science/Grid community which have requirements for very high-speed connectivity and QoS-controlled network access. This proposal is supported by Sun Microsystems, Cisco Systems and UKERNA.

Starts: 01 October 2007 Ends: 30 April 2008 Value (£): 67,624

Scheme: Standard Research

EPSRC Research Topic Cl as sifications:

ICT Networks and Distributed Systems

EPSRC Industrial Sector Cl as sifications:

Communications

-----------------------------------------------------------------------------------------------------------------

EPSRC Reference: GR/T18615/01

Title: Whole System Modelling Of Large-Scale Communication Networks For What-If Evaluation

Principal Investigator: Professor JM Pitts

Other Investigators: Professor DK Arrowsmith

Dr J Schormans

Dr RJ Mondragon

Dr C Phillips

 

Department: Electronic Engineering

Organisation: Queen Mary, University of London

Abstract: Common simulation platforms have played a key role in the design and deployment of wide-area Internet protocols. These protocols are vital to the exchange of electronic information - the life-blood of the modem knowledge-b as ed economy. This project aims to develop new methods for supporting the simulation of large-scale communications networks by using national supercomputing services. Many are as of computational science and engineering research are incre as ingly being focused on the simulation of whole systems, rather than just system components. This project aims to create a new user community that will harness these high-end computing facilities (a) for researching the next generation of networking services and applications, and (b) for what-if evaluation of carrier-scale (i.e. whole system) mobile and broadband network infr as tructure. ' The research will be b as ed around open, extensible simulation technologies (ns-from the Internet community; HLA-from the defence sector) to encourage speedy take-up by other networking research teams. The project will also address the significant methodological challenges inherent in large-scale network simulation. Shared network infr as tructures are known to exhibit complex, highly variable and non -linear patterns of behaviour. The methodological challenges relate to the scale and complexity of network scenarios, the me as urement, analysis and visualisation of results, and the reproducibility and validation of research outcomes for large-scale scenarios. The research will be undertaken by the Network and Service Assurance Laboratory, part of the Communications Research Group at Queen Mary, University of London, in conjunction with national supercomputing services at the University of Manchester.

Starts: 01 October 2004 Ends: 30 September 2007 Value (£): 237,237

Scheme: Standard Research

EPSRC Research Topic Cl as sifications:

ICT Networks and Distributed Systems

EPSRC Industrial Sector Cl as sifications:

Communications