Browse Prior Art Database

Method to create a Cloud "Flood-Plain" for keeping cloud workloads active during maintenance

IP.com Disclosure Number: IPCOM000199374D
Publication Date: 2010-Aug-31
Document File: 2 page(s) / 29K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a method to create a cloud "flood-plain" for keeping cloud workloads active during maintenance.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 1 of 2

Ȉ ˇ ˄

Ȉ ˇ ˄ Ȉ ˇ ˄

Ȉ ˇ ˄

˙

˙ ˙

Disclosed is a method to create an automatic mechanism by which the cloud can assimilate non-cloud members during maintenance periods. These machines act as a cloud "flood-plain", allowing the cloud to overrun its normal definition during times of hardware related failures requiring maintenance without permanently altering cloud membership. The cloud would "draft" whole or partial machines into the cloud based on the needs of applications being evacuated from failing nodes.

    As cloud systems become a more widely implemented platform for production, the maintenance of these systems will become necessary tasks. The cloud is intended to provide a highly available and resilient platform for running these applications. When maintenance needs to be done, it is imperative that these workloads remain available throughout the process.

    Clouds are also intended to be highly utilized. This means that the systems in the cloud should be at capacity as often as possible to maximize the benefit of the cloud environment. To accomplish this, cloud management systems have a software system that optimizes where workloads are placed onto members of the cloud as part of a virtual machine/appliance deployment. These management systems work by identifying machines intended to be part of the cloud, evaluating their current capacity and selecting a location best suited to hosting the new workload. In the event of a hardware failure or maintenance event, all of the workloads (virtual machines) need to be redeployed from the target systems to other members of the cloud. This is typically referred to as evacuation.

    A problem arises that when a system needs to be taken down; not only do the workloads need to be relocated, but the cloud has also lost a subset of its resources. This increases the utilization level of the cloud by twice the capacity of the machine(s) removed. In customer environments, a maintenance event is likely to cause a significant strain on resources. In this scenario, they are often left with a very hard-to-manage manual process of determining what virtual appliances can fit on which machines, possibly degrading the performance of the cloud even further, to facilitate maintenance of the systems. It would be much more palatable to perform maintenance without having to incur this penalty or take additional manual steps.

    The result is that workloads running in the cloud can run with no interruption in the flood-plain while cloud servers are being worked on. Additionally, the cloud workloads and servers do not need to be suspended nor restricted of resources because now additional, temporary resource is available. The method works in two ways:
1) The method scans all hardware for those servers compatible with the cloud workloads. Those servers, once approved by the administrator, are then added to a 'Flood-Plain'.

No changes are

made on those servers, but validation is performed to make sure that the...