The general context that underlies the proposed study is the convergence of telecom and IT as illustrated by services on the Internet and the desire of telcos to develop multiple convergent services accessible from any type of terminals through any access networks. A central challenge here for telcos is to be able to manage end-to-end quality of service in its broader definition (dependability, performance, user experience, …) – even though diversity and heterogeneity are constantly increasing in networks (IP, xDSL, GSM, GPRS, UMTS, ZigBee, Bluetooth…) and terminals (PCs, mobile phones, smart phones, PDA, set-top boxes, communicating objects…).

In this context, network and software infrastructures are omnipresent, complex, and critical. It is observed that exploitation costs (deployment, supervision, bug fixing, maintenance, evolution) are currently moving drastically from costs related to the hardware, development and licensing to costs related to exploitation maintenance and evolution. It appears also quite clearly that a “manual” management of pervasive environments such as machine to machine (M2M), automotive, home networking, or large scale distributed infrastructures such as grids and P2P systems and other overlay networks are in practice, almost impossible. In a sentence, “computing system's complexity appears to be approaching human capability” (J.O. Kephart, D. M. Chess - IBM), even for most skilled administrators.

Autonomic computing is a strategic direction promoted initially by IBM (who coined the term “autonomic computing”) since 2001 and now endorsed by many actors in academia (e.g. Stanford, Berkeley, Carnegie Mellon, Rutgers, Arizona, Imperial College, INRIA) and industry (e.g. Motorola, Huawei, Cisco, Oracle, BEA, HP, INTEL). Autonomic computing and networking aims basically at, as much as possible, automating the management (administration) of network and software infrastructures in order to decrease human interventions and associated costs, enhance dependability and security, and adapt performance to varying workloads. Autonomic network and systems typically exhibit properties such as self-repair, self-optimization, self-protection, self-configuration - collectively known as ‘self-* properties’. This automation of administration procedures are in accordance with the high-level objectives or policies described by the human administrators who can concentrate on high-value tasks while the mundane and monotonous operations are managed (more or less automatically) by technology.

Agile, flexible, adaptable network and software infrastructures and their management have emerged as important matters for telcos with respect to

• huge expected OPEX and CAPEX cost reductions

• more reliable and efficient infrastructures and services

These matters can be potentially addressed by self-management principles; however major challenges and roadblocks seem to appear before adoption of autonomic technology by telcos, e.g:

• The full vision of autonomics is very ambitious. Application domain inside telco infrastructures and services is very large: core networks, access and home networks, mobile networks, middleware and service platforms, services and information systems, M2M, automotive and other vertical applications - just to cite a few. Autonomic technology will generally be disruptive with respect to current telco infrastructures and practices - hence asks the question of the roadmap for smooth adoption by telcos at short, mid and long term.

• Building blocks of autonomics technology are numerous and complex (monitoring probes and probes aggregation, probes date analysis, diagnostic and decision making, planning and execution of reconfiguration actions, context management and data mining, awareness of services situations enabling semantic mediation for transport peering and collaboration etc.). Telcos could appear much more as an integrator of autonomic technology rather than main producers which appear to be rather on the side of network and IT manufacturers (typically Cisco, Motorola, Huawei, IBM) – which asks the question of what could/should be the positioning of telcos so as to adopt autonomic technology while preserving their assets.

What are the main objectives of this Project?

The main objective of the study is to identify autonomics as a disruptive technology and assess its impacts on telcos’ business. More specifically, the study can be used

• as a pre-study for larger and more concerted initiatives by telcos

• lowering the barriers for take up of autonomics by telcos

• reducing the cost of interoperation between autonomics infrastructures from different telcos by planning the profiling of standards

More detailed objectives of this study are to:

• Define the perimeter of the technology and provide shared understanding of concepts and definitions. Discuss the characteristics of autonomics in IT and networks (What are the specificities of autonomic networking compared to mainstream autonomic computing as expressed for instance in IBM’s autonomic manifesto? Does autonomic networking really exist from a scientific/technical point of view or it is just application of autonomic computing in networks?)

• Identify most promising target usage areas of autonomic technology from an operator point of view (access network, core networks, middleware, service platforms, interpersonal services, storage services, information systems, M2M, home networking, enterprise market, etc.)

• Propose a number of use cases for use of the technology and illustrate its applicability to a service platform, in particular for M2M.

• Identify opportunities and risks for telcos in current telecom and IT convergence. Identify strategies for telcos compared to other players (IBM, Oracle, Cisco, Motorola, etc.).

• Identify current scientific/technical approaches by academics suggesting architectures, diagnosis and decision making, dynamic reconfiguration, interference and stability, as well as migration paths from legacy infrastructures.

• Depict the standardisation (and possibly regulation) context. Identify potential areas of standardization and associated standardization bodies.

• Provide current usage and future adoption roadmap by telcos and possibly recommendations on strategic positioning, architectures, technologies, standards and products.