Method and System for Context-aware Non-Interruptive and Seamless System Maintenance in Cloud Environment
Publication Date: 2012-Feb-01
The IP.com Prior Art Database
Disclosed is a system applicable to the maintenance of running cloud instances for cloud service management. The approach allows a user to initiate updates at any time and implements fixes with minimal interruption. The approach incorporates an advanced view of the system through a management portal that shows the logical view and physical view of instances. The system creates cloned instances to avoid interruptions and loss of data during updates.
Page 01 of 4
Method and System for Context -aware Non-Interruptive and Seamless System Maintenance in Cloud Environment
One of the major administrative tasks in cloud management is the maintenance of the running cloud instances. Typically, administrators must update those instances with fixes or upgrade them to next release/version of the operating system (OS) or certain middleware.
A simple, commonly used solution is to manually stop, update, and restart each instance. This solution is very expensive and error-prone. Administrators have to carry out this process out at the same time to make updates. This forces instances to abruptly stop, possibly resulting in the loss of some important data in process. This also affects the performance of the overall system if there are few hundreds or even thousands of instances requiring updates. Prior notification and reserved maintenance time can resolve those issues but it still forces many instances to stop periodically. Other issues include: an instance may not have much choice as far as maintenance is concern; some, but not all, instances may be interested in a certain fixes; instances may not require upgrades at the same time.
Proposed herein is a better way to apply maintenance in which a user initiates updates at any time and implements fixes with minimal interruption.
A cloud environment, takes advantage of having better knowledge of the images and instances, as well as having more flexibility to clone an existing image into an instance. A better solution is to automatically update all the affected instances through the provisioning system. The provisioning system is aware of the list of the instances for affected image, and is able to perform previous manual process automatically through executing scripts (workflows) to control the logic flow of the updates with minimal impact on the services provided by those instances.
In today's cloud implementation, to protect the data integrity, the common practice is to have the data and the application separated into different components and independently managed. That provides the feasibility of having the software and middleware updated without impact on the data itself. The system easily replaces application instances. The end-user really does not care about the physical instance.
The other important fact is that any software management is the responsibility of the cloud rather than the end-user. In that sense, any software or service that wants to be part of the cloud has to be captured as a cloud image and provided through the standard provisioning process. Based on those cloud features, the disclosed solution is applicable in most environment service applications. It might not fit the real-time application like business financial transaction applications.
Note: In this context, image refers to a static image file which can be provisioned to create multiple instances with same software stacks and versions. An instance refers to a virtual machine whi...