Emanics WP7 - Scalable Management

Phase 1

Motivation

Almost every research group, industrial forum or standardization group in the discipline of network and systems management gets, at some point in time, confronted with scalability issues. Over the last decade this aspect of management was addressed in an unstructured way; each management framework or algorithm proposed its own definition of scalability and performed some specific, home-brewed performance evaluations. Although this approach was understandable in the early days of management, it does not provide a sufficient basis to compare the scalability aspects of different frameworks and algorithms. This situation is no longer viable since multiple approaches compete and scalability has become one, of if not the major issue in management. As the database community did a couple of years ago, the network and service management community needs to work together to define a common understanding of scalability, and provide uniform definitions and evaluation criteria to ensure solid and sustained progress on this facet of management research.

Objectives

The objectives of this work package are to:

• Establish a common vision and provide uniform guidelines for scalability research in the area of management
• Collect performance and scalability data on the various management related challenges within the “Broadband for all” theme
• Provide access to performance data of different management frameworks
• Offer test beds for third parties to evaluate the performance of their management frameworks.

Tasks

• T7.1: Collect scalability requirements within the management plane. Such collection will be achieved through interaction with industry at theta-days, at dedicated workshops and in the context of standardization events. This collection can be complemented by the analysis of parameters and traces of real systems (web server logs, MIBs, traces of management data etc.)
• T7.2: Define a common management benchmarking methodology. This methodology will be evaluated within multiple application scenarios and will address all layers of the management plane (device, network, service and business management). To this end common metrics and procedures to measure such metrics will be identified. A potential outcome of this work could be an extension of the IPPM framework, which is currently standardized within the IETF. The virtual lab of EMANICS will used as far as appropriate to make it available to all NoE participants
• T7.3: Perform benchmarking tests on specific management environments and make the benchmarking data available to the community. This work package will generate the data while their publication will be done through work package 2

Partners

The following EMANICS partners are involved in this work package:

• INRIA LORIA (INRIA) (Olivier Festor)
• University of Twente (UT) (Aiko Pras)
• International University of Bremen (IUB) (Juergen Schoenwaelder)
• Poznan Supercomputing and Networking Center (PSNC) (Jan Weglarz)
• University of Surrey (UniS)(George Pavlou)
• University of Pitesti (UPI) (Luminita State)

Deliverables

The deliverables produced by this workpackage can be found here.

Phase 2

Objectives

The objective of this second phase is twofold: firstly, the work package will foster high quality work on producing reports reflecting the current state of the art in the identified work directions. Secondly, mutual and multilateral interactions among the partners will be strongly encouraged and a more flexible and decentralized working framework will be provided.

Description of work

Large scale management concerns all the five functional areas of network management (FCAPS). WP7 will address only three of these areas: security, fault and configuration, the other two being addressed in WP8 and WP9. The work-package is structured among two tasks:

T7.1: Fault, Configuration, and Security operation in large scale environments

Design, model and evaluate algorithms in the three functional areas to cope with large scale systems. The key expected outcomes are:

• To maintain a public inventory and complete high quality state of the art on recent research outcomes related to fault, configuration, and security operation in large scale network management.
• To provide architectural and functional patterns for management frameworks suited in large scale analysis.

T7.2: Modelling and benchmarking

Building up on results from the first phase of WP7, this task will:

• Maintain state of the art documents on methodologies, approaches, results, and research challenges for benchmarking management frameworks.
• Produce both analytical and empirical models of management behaviours.
• Evolve towards effective joint benchmarking campaigns for existing technologies over various support frameworks.

Organisation of the work-package

The work to be addressed in the two themes of WP7 will be organized on a open competitive proposal-based mechanism. Call for collaborative proposals will be launched in two steps. Cooperative proposals including at least two partners will be selected by the work-package leader together with the Execom. Selected proposals will provide for delivery both extensive state of the art on the theme and an in depth presentation of the activity achievements. Reporting is done internally every 3 months by the selected activities. An oral presentation at the WP meetings is done every 6 months. Every work-package partner commits to participate to all WP meetings and to participate in at least one proposal.

Partners

The following EMANICS partners are involved in this work package:

• University of Twente (UT) (Aiko Pras)
• INRIA LORIA (INRIA) (Olivier Festor)
• International University of Bremen (IUB) (Juergen Schoenwaelder)
• Oslo University College (HIO) (Mark Burgess)
• Poznan Supercomputing and Networking Center (PSNC) (Jan Weglarz)
• University of Surrey (UCL)(George Pavlou)
• University of Pitesti (UPI) (Luminita State)

Deliverables

The deliverables produced by this workpackage can be found here.

Phase3

(Large Scale Management)

Objectives

In the final year, the objective of the work package is the collaborative research between the partners in the field of large scale management with the goal to develop, evaluate and report new approaches and solutions in the area.

Description of work

Scalability is one of the major issues in network management. Many management solutions that are available for small networks can not be easily applied to networks consisting of hundreds or thousands of communication entities. The research activities in this work package focus on the security, configuration and monitoring in large-scale environments. Work is structured around the following two tasks (with possible extensions if required):

T7.1: Security in large-scale environments

Activities in this task address the design, implementation and evaluation of new techniques to manage the security in large-scale environments, comprising the detection of intrusions and attack attempts, the secure access to resources, etc.

T7.2: Monitoring and configuration in large-scale environments

Large-scale networks pose several challenges for management, for example the large amount of traffic transported in such networks, the large and variable number of participating network entities and involved parties, etc. Activities in this task will follow different approaches such as distribution and virtualization to allow the effective monitoring and configuration under such conditions.

Organisation of the work-package

The work to be addressed in the two themes of WP7 will be organized on the result of an open call for proposals that was issued at the end of the previous year. Cooperative proposals including at least two partners will be selected by the work-package leader together with the Execom. Selected proposals will provide for delivery an in-depth presentation of the activity achievements.

Reporting is done internally every 3 months by the selected activities. An interim report (D7.5) describing the current activity achievements will be delivered in mid 2009, followed by a conclusive report at the end of the year (D7.6).

Active projects during 2009

Extended Scalable Management of Biometric devices (BioScale II)

Biometric devices emerged to a mature technology and begin to fulfil its promises for physical access control and login to computers. But with today’s device specific software provided by manufacturers, medium and large scale deployments of biometric devices is unfeasible. One of the main challenges of such large scale deployment projects is a scalable distribution system for biometric templates. The objective of the BioScale I and II activity is the development of a scalable management approach for biometric devices.

Flow-based Monitoring and Anomaly Detection (FMAD)

The spread of Gbps networks and the constantly increasing amount of traffic transported by such networks has drawn the attention to scalability aspects of the management of such networks, leading to an interest in flow-based techniques for analysis and anomaly detection in large and/or fast networks. In this activity, we have developed and evaluated several techniques and approaches that allow characterizing and analyzing flow data collected in such large-scale, resp. high-speed networks.

Management of large MANETs (MaMANET)

Mobile ad-hoc networks will play a vital role in future mobile networks. This technique, however, raises new challenges with respect to the management of such a system. Current management approaches address only some of the aspects identified in that field. Thus, this project addresses the following topics that have been the areas of interest in the MaMANET activity: 1)requirements analysis; 2) define a comprehensive management architecture for MANETs; 3) detailed part 1: propose service search; 4) detailed part 2: resource management of disruption tolerant networks

Security Management Infrastructure (SeMI)

An end-to-end secure communication in an inter-domain and heterogeneous environment is a key yet complex factor. One of the most important management challenges is to assure a predefined security level over multi provider environments with dedicated communication infrastructures, security mechanisms, processes and policies. Thus, it is necessary to overcome management islands, mostly realized by commercial providers, and to work towards a common security management infrastructure (SMI) to get a common understanding of the security services and mechanisms to realize the requested level of security.

Virtualization Monitoring (VirtMon)

Virtualization is called to be a key technology in the future Internet. Nevertheless, this paradigm is not exempt of penalties because the virtualization process imposes overheads that slowdown the network processes. Then, a careful analysis has to be done before applying the technique to get an appropriate trade-off between benefits and drawbacks. In addition, a key requirement is to find appropriate monitoring mechanisms for the virtualized resources and networks. The VirtMon project is addressing these challenges.

Partners

The following EMANICS partners are involved in this work package:

• University of Twente (UT) (Aiko Pras)
• INRIA LORIA (INRIA) (Olivier Festor)
• Jacobs University Bremen (JUB) (Juergen Schoenwaelder)
• University Politecnica de Catalunya (UPC) (Joan Serrat)
• University of the Federal Armed Forces Munich (CETIM) (Gabi Dreo-Rodosek)
• University of Zurich (UNIZH) (Burkhard Stiller)
• Poznan Supercomputing and Networking Center (PSNC) (Jan Weglarz)
• University of Surrey (UCL)(George Pavlou)

Deliverables

The deliverables produced by this workpackage can be found here.