Content-Based Scheduling of Virtual Machines (VMs) in the Cloud

Abstract

Organizations of all sizes are shifting their IT infrastructures to the cloud because of its cost efficiency and convenience. Because of the on-demand nature of the Infrastructure as a Service (IaaS) clouds, hundreds of thousands of virtual machines (VMs) may be deployed and terminated in a single large cloud data center each day. In this paper, we propose a content-based scheduling algorithm for the placement of VMs in data centers. We take advantage of the fact that it is possible to find identical disk blocks in different VM disk images with similar operating systems by scheduling VMs with high content similarity on the same hosts. That allows us to reduce the amount of data transferred when deploying a VM on a destination host. In this paper, we first present our study of content similarity between different VMs, based on a large set of VMs with different operating systems that represent the majority of popular operating systems in use today. Our analysis shows that content similarity between VMs with the same operating system and close version numbers (e.g., Ubuntu 12.04 vs. Ubuntu 11.10) can be as high as 60%. We also show that there is close to zero content similarity between VMs with different operating systems. Second, based on the above results, we designed a content-based scheduling algorithm that lowers the network traffic associated with transfer of VM disk images inside data centers. Our experimental results show that the amount of data transfer associated with deployment of VMs and transfer of virtual disk images can be lowered by more than 70%, resulting in significant savings in data center network utilization and congestion.

A Cross-Entropy-Based Admission Control Optimization Approach for Heterogeneous Virtual Machine Placement in Public Clouds

“…Bazarbayev et al [9] propose a content-based scheduling algorithm for VM placement in cloud data centers, by utilizing similarity between VM disk images. Their approach can lower the network traffic during the transfers of VMs between racks in data centers and results in significant savings in data center network utilization and congestion.…”

Optimal Energy Efficient Scheduling based on VM Energy Consumption in Cloud

“…. Sobir Bazarbayev et al [13] in their work, authors discussed the Infrastructure as a Service (IaaS) clouds, provides on demand nature of services. In the large cloud data centre hundreds of thousands of virtual machines can be deployed and terminated.…”

“…Next, the VMs are classified into three types based on the resource usage as CPU intensive, I/O intensive or Network intensive [12]. Finally the VMs are scheduled onto the physical servers [13] taking into account the resource utilization and energy consumption of the server such that server utilization is increased but energy consumption is minimal [14], [15]. The technique was simulated on CloudSim [16] and experimental results show the algorithm effectiveness of the proposed technique.…”

A Deduplication-Enabled P2P Protocol for VM Image Distribution

“…Unlike those efforts where primary focus is on the adoption itself of P2P BitTorrent-like protocols for data dissemination in the cloud, our dedup-enabled VM provisioning scheme explores a possibility of deduplication combined with peerto-peer content delivery protocols. Our approach is motivated by the recent studies of VM images for representative cloud offerings that reported a significant degree of similarity at image and block level [5], [6]. Figure 1 illustrates our approach of deduplicated VM image distribution in a peer-to-peer fashion.…”

“…Various delivery schemes explored can be categorized into either C/S-or P2P-based protocols. Many prominent efforts in the former group pursue data efficiency by avoiding or minimizing redundant data access or transfer [6], [11]. One notable approach is to perform deduplication with the help of block translation maps generated beforehand when VM images were created, at a higher level than the usual memory or storage device layer [11].…”