Kari Seppänen

Enabling wireless backhauling for next generation mmWave networks

This paper presents some key findings w.r.t. the Radio Resource Management (RRM) in wireless Backhaul (BH) of mmWave networks. First, the envisioned design of mmWave backhaul architecture is outlined highlighting the most important newly needed functional blocks and in what they differ from a non-mmWave architectures. Next, the challenges and functionality of RRM techniques are discussed, focusing on Routing and Link scheduling algorithms in such BH architecture. Furthermore different possible interactions between RRM functions are explored. Finally, preliminary analytical and experimental study on the performance of different link scheduling and routing functions for mmWave backhauling are provided, highlighting in particular the impact of traffic load and dynamic route selection on BH End to End delay.

Self-optimizing last-mile backhaul network for 5G small cells

We describe a novel wireless backhaul solution, called Self-Optimizing Wireless Mesh Network (SWMN), which is targeted for interconnecting 5G small cell nodes cost-efficiently and smartly. SWMN is not only highly resilient and scalable, but it is also equipped with advanced Self-Organizing Network (SON) capabilities offering autonomous network configuration and traffic engineering features. The key aspect of SWMN is an innovative network intelligence engine coupled with beam steerable pencil beam millimeter wave (mmW) high capacity backhaul radio.

Multipath routing for mmWave WMN backhaul

Static Wireless Mesh Networks (WMN) based on high capacity pencil-beam millimeter wave (mmWave) links offer an attractive solution for small cell backhauling in 5G networks. However, the delay targets in 5G are such that it is quite challenging to develop adequate link scheduling and routing mechanisms. Furthermore, the susceptibility of mmWave radio links to adverse weather conditions and other obstructions in the line-of-sight line makes it necessary that routing is able to cope with frequent link degradations and failures. We propose an approach that divides the problem into two parts - finding first a suitable set of so called primary paths and then optimizing the link schedules for those paths. In this paper, we present a routing algorithm that produces the required primary paths and several backup paths for all source-destination pairs. Furthermore, we describe a hierarchical fault recovery system that includes various protection and restoration methods that are acting in different time scales to ensure fast reaction to failures and then optimal adaptation to the changed situation.

Fair queueing for mmWave WMN backhaul

Static Wireless Mesh Networks (WMN) based on high capacity pencil-beam millimeter wave (mmWave) links offer an attractive solution for small cell backhauling in 5G networks. However, the delay targets in 5G are such that it is quite challenging to develop adequate link scheduling and routing mechanisms especially for a system composed of backhaul nodes with a single TDD radio. We have proposed earlier a scheme that is based on cyclic semipermanent link schedules and multiple spanning tree based multipath routing. This scheme enables multihop backhaul connections with low delays and fast fault recovery. However, so far this scheme has been analyzed only by using end-to-end delay estimates based solely on link schedules. The potential effects of different queueing disciplines have not been studied in detail. In this paper, we will show how common output queueing systems can have a harmful interplay with the cyclic link schedules that contradicts our primary path optimization targets. We will also suggest a few simple modifications to existing queueing solutions that resolve these problems at WMN node level.

Integrating WMN Based Mobile Backhaul with SDN Control

Resilient high capacity and low delay millimeter wave wireless mesh networks (WMNs) can provide suitable backhaul connections for future broadband mobile networks. The WMN solution is best suited in cases where base stations are installed in locations without optical fiber connection to transport network, e.g., small-cell deployment to hot spots in dense urban areas. Recently software defined network (SDN) concept has become popular in many networking areas including mobile networks. One of the key promises is to provide an efficient way for network operators to extend and create new services. As the whole network is controlled by a single central entity that is based on software code, it would be easy to make large scale network upgrades without need to wait that updates are available for all network elements (NEs). There is, however, a clear conflict between SDN ideas and WMN operation. The performance and reliability of the latter one is heavily depended on fast local reactions to, e.g., link degradations. Centralized control would introduce longer delays in reactions. In this paper, we are proposing a concept which solves these problems and allows for combining the best features of both WMN and SDN.

Link scheduling for mmWave WMN backhaul

A new link scheduling approach for millimeter wave wireless mesh network is presented. The proposed link scheduling solution is designed to work together with a multiple overlapping spanning trees-based routing schemes on a mesh network topology. The link schedule computation will minimize the end-to-end delay along a subset of shortest-hop paths from non-gateway nodes to a gateway node of the mesh, where each path is obtained from one of the spanning trees. The link schedule itself consists of cyclic repetition of a vector of transmission sets, where a transmission set is a maximal collection of links that can be active simultaneously. Simultaneous activity of the links is limited by radio interference and by shared resources. Algorithms for the computation of the link schedule are provided and computational issues of the algorithms are discussed.