Service Optimization

Traditionally, Internet-based services have had few options on how to optimize communication between involved parties. This was previously not a problem as most services had one major requirement, high throughput, which also was considered by nearly all components of Internet hardware and software. Recent years have, however, seen a major increase in both services with other requirements and in technologies, used to access the Internet, with significantly different characteristics. For example, there is several new initiatives to consider network latency instead of throughput, as the crucial performance metric as most web-services are dependent on low latency rather than high throughput. Furthermore, mobile devices like smart phones and tablets, that uses 3G/4G based Internet access technologies, have different characteristics than ordinary computers, e.g. the need to conserve energy, multiple network interfaces that can be used simultaneously, and so on.

The key challenge of this project is to design and evaluate new architectural solutions, resource allocation methods, and protocols that can support heterogeneous service requirements simultaneously. We will develop such solutions by considering more expressive applications and transport services in user terminals, enhanced resource allocation methods in routers, and how to best utilize multiple interfaces in mobile devices.

To accomplish such mobile service optimization we will consider and update current research in these areas. For instance, the IETF has recently started initiatives that address the need to allow applications to communicate their specific requirements to other networking components, which then are supposed to treat the application traffic accordingly. The current initiatives cover both how to communicate such requirements to the local system [84] as well as to other network components, such as home gateways or operator networks. To make use of this information the local system as well as operator networks and home gateways need to revise how they allocate resources. For instance, if a certain application requires real-time communication a gateway might have to enforce a certain active queue management (AQM) strategy or optimize its operation in some other way. It is also possible to optimize resource allocation with extended information from network operators. For example, by conducting traffic steering and control both from the perspective of the mobile devices that typically have multiple air interfaces, the home gateways that serve as an aggregation point for multiple in-home devices towards the operator network and within the network elements in order to direct the traffic towards network elements that allow to make service optimization decisions. For flexible and efficient traffic steering, SDNis an interesting candidate to provide network operators with a toolbox that allows to exercise control and optimize their network dynamically and efficiently. Although such an approach could be extremely efficient in order to optimize, we will also consider scenarios where the mobile terminal itself can make optimization decisions related to its multiple interfaces.