Institutional Repository
Technical University of Crete
Technical University of Crete
EN
|
EL
| URI: | http://purl.tuc.gr/dl/dias/7CDBE299-9F0A-4810-9073-431940F4678B | ||
| Year | 2017 | ||
| Type of Item | Diploma Work | ||
| License |
|
||
| Bibliographic Citation | Dimitrios Kargatzis, "Virtual machine deployment and migration on heterogeneous cloud platforms", Diploma Work, School of Electrical and Computer Engineering, Technical University of Crete, Chania, Greece, 2017 https://doi.org/10.26233/heallink.tuc.67729 | ||
| Appears in Collections |
Cloud computing offers an innovative business model for organizations to adopt ITservices at a reduced cost with increased reliability and scalability. Adopting a cloudsolution means binding to a specific platform and vendor, using specific protocols,standards and tools of the cloud and finally, running into a vendor lock-in situation.The fear of vendor lock-in is often cited as a major impediment to cloud serviceadoption. If the provider decides to raise its prices or change its security policies, thecustomer may have to consider to move his workloads to another provider. In thiswork we focus on the automatic migration requirements in Openstack systems and asa use case we present a mechanism for Virtual Machine (VM) migration (or of theirrunning instances) between Openstack and KVM virtualization and another cloudplatform that runs a different Virtualization engine (Stratogen and VMware is ourcase). The standards approach is to freeze the running instance of a VM and restart itunder the new environment. We examined the requirements for successful migrationand the conclusion is that migration is not always fully automatic as it might need lotsof parameter tuning depending on differences in the type of virtualization enginesused at source and target environments (a task that can be performed by a specializedexpert). To alleviate the requirement of human intervention in the loop, we suggestusing containers as the underlying virtualization technology. We propose amechanism that implements migration using containers in a few steps and we run aseries of experiments to show proof of concept. As a conclusion, the later technologyproved to be feasible and more promising although it is still not fully supported byinfrastructure (operating system) tools that allow migration independent of the state ofthe underlying operating system kernel at the time of transfer.
Copyright © DIAS 2013
