Supporting migration of Virtual machines running Multi-tier applications in Computing clouds that span multiple IP address ranges
Publication Date: 2010-Aug-19
The IP.com Prior Art Database
This invention solves the problem of moving Virtual Machines from one IP subnet to another, which may occcur in large cloud environments. The proposed solution is method to allow a group of related Virtual Machines in a cloud to continue intercommunicating when one member of the group has its IP address changed during a Virtual Machine migration. It can be implemented as firmware or as software. For example, it can be implemented in the Virtual Machine hypervisor or as a separate Virtual Machine Migration Manager software.
Supporting migration of Virtual machines running Multi -tier applications in Computing clouds that span multiple IP address ranges
1. Background: From a user and business perspective, cloud computing is a delivery and consumption model for IT services, where the user sees only the service, and does not need to know anything about the underlying technology or implementation.
There are broadly two types of clouds that users can use: Software as a Service clouds and Platform infrastructure clouds. Each delivers IT services and resources at different levels of the technology stack, and are targeted at different consumers. Software as a Service clouds deliver software usage to end-users, and many providers sells these services as hosted software that can be used over the public Internet. Platform infrastructure clouds deliver middleware, server, storage and network resources to IT and software development teams. Platform infrastructure clouds are commonly built within an enterprise, and serve the needs of the Software Engineers and IT operations personnel within the organization.
Dynamic virtualization is a technology that enables the creation of effective platform infrastructure clouds. It allows middleware and virtual machines to be created, and provisioned on demand. It also allows virtual machines to be moved from one physical server to another, to allow migration to servers with higher computing capacity or to allow planned server maintenance to take place. One of the features of dynamic virtualization is that it allows a virtual machine to maintain its IP address even as it moves from one physical machine to another. However, present technology requires that both virtual machines must be located within the same Layer 3 IP subnet for this to happen. This is because of the nature of IP routing mechanisms, which are inherently tied to physical geography. In some cases, the requirement is more restrictive in that both machines must be located within the same Layer 2 broadcast domain.
This presents a limitation on the use of dynamic virtualization for building platform infrastructure clouds. In large production clouds, it is likely that virtual machines will need to be moved between physical servers that are in different Layer
3 IP subnets. This can occur for a number of reasons. For example, a higher capacity machine may only be available in a different geographical location which is in a different IP subnet and also a different broadcast domain. As another example, a primary server may be undergoing a catastrophic equipment failure, and it may be necessary to migrate a virtual machine from it to another physical server located in a disaster recovery center located in another geographical location.
In these cases, the migrated virtual machine will not be able to retain its original IP address. (This is not the only reason why a migrated virtual machine cannot retain its IP address. An IP address change for example, could be required...