Ph.D Thesis


Ph.D StudentFairstein Yaron
SubjectDynamicity and Multi-Commodity in Networks
DepartmentDepartment of Computer Science
Supervisors PROF. Joseph Naor
PROF. Dan Raz
Full Thesis textFull thesis text - English Version


Abstract

The new network generations offer high bandwidth low-latency communication that provides better services and allows the creation of new businesses. This is enabled in part by the introduction of new emerging network paradigms such as Network Function Virtualization (NFV) and Multi-access Edge Computing (MEC).


The main idea behind NFV is decoupling functionality from hardware, leading to agile networks where service location may change swiftly. MEC, on the other hand, allows to move services from large centralized data centers to the edge of the network, where resource limitations are offset by the low latency that results from short physical distance to the clients.


However, in order to utilize these paradigms and tap into their potential, one needs to deploy new resource allocation algorithms. We identify two main categories of problems that should be addressed in this context: First, multi-commodity, which refers to dividing the limited resources between the available commodities (demanded by users). Second, dynamicity, which considers the dynamic nature of networks and the workload.


In this thesis we provide several algorithmic solutions which fall into at least one of those categories. For each problem domain we study, we define a rigorous model and present an algorithmic solution for the specific problem.  We provide analytically proven performance bounds for these algorithms that are compared to the relevant lower bound.   For some of the problems we also present a thorough performance evaluation via extensive simulation, indicating their advantage over other solutions in realistic scenarios.