Efficient Routing and Bandwidth Assignment for Inter-Data-Center Live Virtual-Machine Migrations

Abstract

Using virtualization, service providers can create an abstraction of the physical servers hosted in their data centers, and run their services directly on these abstract entities, called Virtual Machines (VMs). As a direct results of service virtualization, services hosted on VMs can be migrated from a data center to another. VM migration can be performed without or with a minimal service interruption and in this case we talk about live VM migration. The main issue of live VM migration is that memory pages are modified during the process. Due to this fact, same memory pages might be transferred several times, thus increasing both the migration duration and the amount of network resources occupied by the VM migration process. Considering the increasing amount of VM migrations carried on in inter data center networks, efficient strategies for live-VMmigrations bandwidth provisioning are needed. In this paper, we propose and compare various novel Routing and Bandwidth Assignment (RBA) algorithm for the live migration of VMs in a distributed DC infrastructure, under dynamic traffic conditions. We find that assigning a bandwidth to the VM migration request in proportion to the path length with the objective of minimizing the amount of resources occupied by the migration request could improve the network performance. In addition, we present a comparative analysis of the serial and the parallel migration of multiple VMs. We quantify the impact of the migration strategy (serial versus parallel) on blocking probability, migration duration, downtime and resource occupation.

Online Virtual Machine Evacuation for Disaster Resilience in Inter-Data Center Networks

“…We used the model in [12] to obtain these curves, which will be indicated as function-points curves (see Ref. [13] for a detailed explanation). On one hand, we note that, as bandwidth increases, the advantage of online migration in terms of total migration duration is reduced.…”

Resilience of communication networks to random failures and disasters: An optimization perspective

“…It can be achieved using either proactive or reactive set up of the alternate communication paths, such as, for example, backup paths being disjoint with the affected working paths . As resilience commonly involves redundancy, other ways to improve the performance of communication networks after failures require careful planning of the network architecture itself with a particular focus on the increased connectivity (for instance by the addition of network links) or replication of servers/data centers . Also, the respective measures of network resilience can help evaluate the resilience of given network architecture.…”

Efficient Online Virtual Machines Migration for Alert-Based Disaster Resilience

“…In other words, the lower bound for the migration bandwidth is the VM dirtying rate, D. However, in our evaluations, we assume, as in Ref. [18], that the minimum migration bandwidth possible is B min = 1.2 • D. Similarly, we set a limit on the maximum migration bandwidth B min we can allocate when ∆T mi g ∆B ≤ 0.1%, i.e., when the advantage in terms of migration duration is very low. .…”

“…We used the model in [6] to obtain these curves, which in the following will be referenced to as function-points curves (refer to Ref. [18] for a detailed explanation). On one hand, we note that as the provisioned bandwidth increases, the advantage in terms of total migration duration is reduced.…”