Networks could self-organise sooner than we think

“In 2003, anyone searching for the term ‘autonomic communication’ on Google would have found two references. As of November this year, there are almost twenty one thousand,” says Mikhail Smirnov of the Fraunhofer Institute for Open Communication Systems (FOKUS) in Berlin.

Professor Smirnov coordinates the ACCA project, a Future and Emerging Technologies (FET) initiative that ended in September 2006. He and his colleagues have played a large role in raising awareness of this new field, and they are rightly proud of the phenomenal progress of autonomic communication. Smirnov points to a dozen journals on the subject, multiple IEEE and IFIP conferences, four EU integrated projects and several similar projects set up by leading European telecom companies.

The partners in ACCA were responsible for setting a research roadmap, rising awareness and bringing together relevant but disjoint efforts within related initiatives. Though some of these initiatives focus on work unlikely to generate market-ready technology for some years, others are linked to more market-oriented research, including several projects under the IST programme – Ambient Networks, E2R and others.

“Our main achievement was getting the Autonomic Communication Forum (ACF) up and running,” notes Smirnov. Launched in Brussels in late 2004, the ACF already has more than 230 members.

The forum is building a global research community for autonomic communications, and will eventually become an industry standardisation body. The body is built on the premise that a new form of networking – one that is self-organising, self-managing, context-aware and autonomous – is the best way to handle the internet’s growing complexity and the demands placed on it.

Smirnov underlines the fact that the project was deliberately extended to allow the team to stay together for Autonomic Networking 2006, a conference held in Paris from 25-29 September 2006. The project team took charge of an ACF session, which resulted in ACF decision to respond to any future calls under the Seventh Framework Programme (FP7) in a coordinated manner.

Asked for a brief definition of autonomic communication, Smirnov replies “self-management”. But he notes that the term encompasses a host of other technologies, all of which begin with the word “self”. Among them are self-awareness, self-configuration, self-healing and self-implementation.

“Together these technologies allow a network to automatically address many of today’s internet challenges,” says Smirnov. They could, for instance, help networks to guarantee online identity – thus tackling the lack of security and trust still associated with web commerce.

Autonomic communication could also benefit next-generation networks. Smirnov notes how internet service providers (ISPs) struggle to manage wholesale interfaces – the interface for data traffic between themselves: “Corporate and retail customers stay within a network, so ISPs can prioritise the way they handle this traffic and charge more. But when traffic crosses wholesale interfaces, prioritisation mechanisms break down and ISPs find it harder to charge for service differentiation. Self-management could change that.”

Future networking technologies will also have to solve many internet problems at once. Autonomic communications represent a new design approach to such problems. They offer manageability and therefore address the question of complexity – which Smirnov calls “the big challenge”, especially for telecom companies.

“Top of the pyramid of technologies is self-management. We believe this is the solution to networking complexity,” Smirnov says. “Not least because it is unaffected by hardware and software developments.”

He adds that although self-management makes a system apparently simpler for an administrator, below the surface the complexity is still there. “It’s not about putting intelligence into hardware, but making networks behave intelligently without human intervention.”

One of the companies attracted by the project was Gingko Networks – a start-up company in autonomic communication. “Its multi-agent approach works on network functionality,” says Smirnov. This system calls on several autonomic communication features, and is already in use by France Telecom.

He emphasises, however, that the project partners were involved in coordination rather than technical work: “We made sure researchers were on the right track, and that projects were building a consensus on a roadmap for this networking solution – one example of which is the Self-Management Technology Platform.”

Smirnov expects autonomic communication to have a major impact within five years. “Companies like my own are working on related technologies, systems and services. The look and feel of communication devices will change, facilitating communications with one’s community.”

“Today, these communications are via an ISP or gateway – in future people will be connected directly to their contacts,” says Smirnov. Ultimately, he adds, the telecom business model needs reinventing to enable better links between users, content and service.