Browse Prior Art Database

Dynamic transfer of cloud workloads to minimise latency

IP.com Disclosure Number: IPCOM000239935D
Publication Date: 2014-Dec-16
Document File: 1 page(s) / 19K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a method to dynamically manage the physical location of Cloud workloads. Typically when deploying applications to Cloud infrastructure the physical data centre location is decided upfront, based on fairly static knowledge about where the majority of customers are for example. The alternative is not to fix the location upfront but to automatically move live workloads geographically to meet business requirements, such as latency targets or cost.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 54% of the total text.

Page 01 of 1

Dynamic transfer of cloud workloads to minimise latency

Some Cloud workloads are sensitive to latency. For example trading applications, games or virtual worlds, operational master data management hubs, etc.. Cloud providers, for example Amazon and IBM's SoftLayer, typically provide data centres in multiple geographic locations and customers would choose one or more of these locations to host their workloads taking any specific latency requirements into account. This approach is not flexible and assumes that latency requirements are relatively fixed.

    In the future Cloud platforms could provide facilities to dynamically move live workloads between data centres based on various business requirements, for example to minimise latency for the majority of active users.

    For example, a global company may have a cloud hosted operational MDM system which is used by customer facing applications. To ensure that those applications are responsive for all their customers, they would currently need to use multiple data centres in every region/country where they have customers.

    In a similar way to 'follow-the-sun' technical support, it would be more efficient to move the MDM workload to the data centre nearest their most active customers throughout the day. Existing scaling techniques could already be used to automatically increase or reduce application instances in each data centre, and in the future it may be possible to move live virtual machines over large distances. Combining various strategies with new APIs tools to manage the geographic location being used for applications in the Cloud would allow companies to reduce latency for their customers more efficiently.

    For example, in the...