Vehicles are an important source of data, with embedded sensors including built-in cameras. Their data can be processed close to where it was generated, using computation and network resources on collaborating vehicles. Processing at the network edge helps to reduce end-to-end service latencies. However, such mobile virtual resource pools require effective management, especially if those resources are to be leased to third-party service providers. We address the problem of service placement on moving, connected, vehicle nodes (V2V) supported by roadside infrastructure (V2I). We formulate a distributed service model and optimize the placement of services, subject to constraints related to node resource capacity, link capacity, distributed application deployment (full deployment, anti-collocation and adjacency constraint) and vehicle mobility.

SLAs are an integral part of all modern service provisioning operations. They have been a topic of discussion, research and development for many years but still the norm is the use of rigid, complex and not easy to automate Service Level Agreements. In this paper we are presenting a service framework that is leveraging the OCCI specification in order to facilitate standardized SLA provisioning and violation detection. This SLA as a Service (SLAaaS) offering is provided as an open source framework to any Service Provider that wants to efficiently enhance his infrastructure with SLA support.

This paper describes a generic and fast classifier that uses a binary CMM (Correlation Matrix Memory) neural network for storing and matching a large amount of patterns efficiently, and a k-NN rule for classification. To meet CMM input requirements, a robust encoding method is proposed to convert numerical inputs into binary ones with the maximally achievable uniformity. To reduce the execution bottleneck, a hardware implementation of the CMM is described, which shows the network with on-board training and testing operates at over 200 times the speed of a current mid-range workstation, and is scaleable to very large problems. The CMM classifier has been tested on several benchmarks and, comparing with a simple k-NN classifier, it gave less than 1 % lower accuracy and over 4 and 12 times speed-ups in software and hardware respectively. 1