Distributed Server Migration for Scalable Internet Service Deployment

Abstract

The effectiveness of service provisioning in largescale networks is highly dependent on the number and location of service facilities deployed at various hosts. The classical, centralized approach to determining the latter would amount to formulating and solving the uncapacitated k-median (UKM) problem (if the requested number of facilities is fixed), or the uncapacitated facility location (UFL) problem (if the number of facilities is also to be optimized). Clearly, such centralized approaches require knowledge of global topological and demand information, and thus do not scale and are not practical for large networks. The key question posed and answered in this paper is the following: "How can we determine in a distributed and scalable manner the number and location of service facilities?" We propose an innovative approach in which topology and demand information is limited to neighborhoods, or balls of small radius around selected facilities, whereas demand information is captured implicitly for the remaining (remote) clients outside these neighborhoods, by mapping them to clients on the edge of the neighborhood; the ball radius regulates the trade-off between scalability and performance. We develop a scalable, distributed approach that answers our key question through an iterative reoptimization of the location and the number of facilities within such balls. We show that even for small values of the radius (1 or 2), our distributed approach achieves performance under various synthetic and real Internet topologies and workloads that is comparable to that of optimal, centralized approaches requiring full topology and demand information. I. INTRODUCTION Motivation: Imagine a large-scale bandwidth/processingintensive service such as the real-time distribution of software updates and patches [2], a distributed data-center [3], a cloud computing platform [4], [5], etc. Such services must cope with the typically voluminous and bursty demand-both in terms of overall load and geographical distribution of the sources of demand-due to recently observed flash-crowd phenomena. To deploy such services, decisions must be made on: (1) the location, and optionally, (2) the number of nodes (or

Distributed Placement of Autonomic Internet Services

“…Even closer to our work is the paper of Smaragdakis et al [26] that proposes the R-ball heuristic. They reduce the original k-median problem in multiple smaller-scale 1-median problems solved within an area of R-hops from the current location of each service facility.…”

“…One example is the R-ball heuristic used in [26], where the search for a better service host is a priori bounded within the r-hop distance neighborhood of the node each time hosting the service 7 . On the contrary, cDSMA invests more effort and intelligence in selecting the service host candidates.…”

“…This task may be carried out by any efficient flooding scheme requiring O(|E|) messages and O(D) time, where D is the network diameter. Note that this step is sine qua non for all protocol instances real- izing distributed service placement under dynamic demand patterns (e.g., [26], [27]). …”

Flow Clustering Based Efficient Consolidated Middlebox Positioning Approach for SDN/NFV-Enabled Network

“…These requirements are not practical for large networks. A more effective approach is the distributed method as proposed in [10]. Generally, the network will be divided into small r-ball regions or r hops away from the facility.…”

Optimal Service Placement in Distributed Networks

“…The second change has to do with the proliferation and "miniaturization" or the network elements, as well as network users. The term "miniaturization" may be used here to capture the fact that the traditionally heavy network elements (routers) are increasingly being supplemented by lighter network elements that are contributed by (until recently) traditional network users; these users are becoming powerful enough to engage in ad hoc networking and contribute to the networking infrastructure [3]. For example, the increasing contribution to the networking infrastructure of numerous small (traditionally) usernodes is already materializing in lastmile networking and is expected to dominate soon (e.g., home owner based WLAN network service access).…”