The dependability of ICT systems is vital for today's society. However, operational systems are not fault free. Providers and customers have to define clear availability requirements and penalties on the delivered services by using SLAs. Fulfilling the stipulated availability may be expensive. The lack of mechanisms that allow a fine control of the SLA risk may lead to over-dimension the provided resources. Therefore, a relevant question for ICT service providers is: How to guarantee the SLA availability in a cost efficient way? This paper studies how to combine different fault tolerant techniques with different costs and properties, in order to economically fulfill a given SLA requirement. GEARSHIFT is a mechanism that enables ICT providers to set the fault tolerance technique (gear ratio) needed, depending on the current service conditions and requirements. We illustrate how to use the proposed model in a backbone network scenario, using measurements from a production national network. Finally, we show that the total costs of delivering an ICT service follow a simple convex function, which allows an easy selection of the optimal risk by tuning properly the combination of fault tolerant techniques.

This paper presents an analysis of measurements on a global route between Norway and China. Measurement statistics was collected every 10 milliseconds during a three-month period in October-December 2009. We provide a detailed characterization of observed loss and delay patterns especially focusing on downtime periods on the sub-second time scale (from 50 milliseconds to 1 second long). Generally, these downtimes are more difficult to measure, but they may also have a noticeable negative effect on quality of real-time services and multimedia applications. The paper investigates causes for all these events and finds that the plausible causes most of observed communication problems in the given study are either link or router related (level 2 or 3). Packet loss events caused by congestion could also be distinguished, but they happened very rarely. Using the collected data set, we also study several other questions related to routing and network availability modeling.

Future user controlled development of telecommunication services combined with powerful terminal equipment results in many heterogenous services running in a peer-to-peer execution environment. Locating a desired service in such an environment is challenging. We present a swarm based optimization algorithm which is capable of finding paths of resources in a complex network environment. The algorithm is fully distributed and may be implemented using simple ant-like mobile agents. On the contrary to existing localization mechanisms for peer-to-peer systems the algorithm considers all accessed resources between (and including) the client side and server side when a resource path is evaluated. Scalability is achieved by making agents cooperate during search when they have overlapping search profiles. Results from simulations are promising. The expected cooperative behavior is shown to be present, i.e. a set of near optimal resource paths conforming to a set of different but overlapping search profiles may be found with improved performance.

Telecommunication network owners and operators have for half a century been well aware of the potential loss of revenue if a major trunk is damaged, thus dependability at high cost has been implemented. A simple, effective and common dependability scheme is 1:1 protection with 100% capacity redundancy in the network. A growing number of applications in need of dependable connections with specific requirements to bandwidth and delay have started using the internet (which only provides best effort transport) as their base communication service. In this paper we adopt the 1:1 protection scheme and incorporate it as part of a routing system applicable for internet infrastructures. 100% capacity redundancy is no longer required. A distributed stochastic path finding (routing) algorithm based on Rubinstein's Cross Entropy method for combinatorial optimisation is presented. Early results from Monte Carlo simulations indeed indicate that the algorithm is capable of finding pairs of independent primary and backup paths satisfying specific bandwidth constraints.

In this paper we examine what failure semantics are desirable for services provided by a mobile agent system (MAS) when assuming the MAS to be part of a network fault management system. We also present results from an evaluation project where the failure semantics of state of the art MAS were examined.