Font Size: a A A

Resilient Hypervisor Based Proactive- Reactive Recovery On A Single Host "Double Layer Clustering Approach"

Posted on:2016-12-01Degree:MasterType:Thesis
Country:ChinaCandidate:L D KeFull Text:PDF
GTID:2308330461461470Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
The advent of virtual machine monitor (VMM), namely Hypervisor has impacted the virtualization world and brought up a possibility to run one or more virtual machines with a variety of operating systems that share the same virtualized hardware resources. Thus, this approach has been used to develop a new way of protecting and maintaining the availability of information so to build a trusted computing environment such as the proactive recovery method that allows the replicated instances to tolerate an unlimited number of malicious faults by periodical rejuvenation of replicas. And with the computing power increasing every 18 months, now with the current hardware available on the market, we can make a single host that can contain multiple clusters of virtual machines.Our approach consists of creating a double layer cluster system that will have each cluster running 3f+1 instances(VM) of the same operating system different from other clusters on a single host, and uses a dispatcher to ensure the random access to the instance, so to increase the resiliency of the system. This approach does not allow communication between the replicas or instances instead the replicas manager is in charge of the communication this is to avoid the spread of infection in case of an attack and the replicas clock is maintained by the replicas manager. The resilience of our approach is presented in three steps that consist of the dispatcher which use an algorithm to obfuscate ports and IP addresses of the online instances. While this is not a replacement for cryptographic methods, the described port number randomization algorithms provide improved security/obfuscation with very little effort and without any key management overhead. The system is maintained through the clusters’ manager which the administrator can access to manually switch off a cluster so to maintain it, and the administrator can check the status of the online replicas via a monitoring system.
Keywords/Search Tags:Virtualization, Proactive Reactive Recovery, Cluster, Intrusion Tolerance, Availability, Diversity
PDF Full Text Request
Related items