IBM devises cloud data centre disaster planning
Big Blue sets sights on finding ways to minimise data centre disruption because of natural disasters.
IBM Labs is working with students at a college in New York State to build software-defined networking (SDN) technology that will allow organisations to quickly reconfigure networks and move data and applications in the event of a natural disaster.
It is hoped the innovation could help prevent disruptions in voice and data communications services caused by hurricanes and other natural disasters.
The cloud-based rapid re-provisioning application, which would move data and applications, comes in response to last year's Hurricane Sandy, which left many data centre operators unable to move resources at short notice.
The invention could cut re-provisioning times from days to minutes, IBM claims. It would also avoid costly network disruptions and outages.
SDN enables data centre operators to more efficiently control data flows within physical and virtual networks.
The SDN advancement IBM and Marist College are testing will enable an IT professional to remotely access and make changes to network resources via a wireless device and open source network controller developed by Marist.
Zachary Meath, a student at the College, said in a blog post the aim of the project was to devise a way to re-provision a network “in a matter of minutes, not days or weeks, which is currently the norm.”
The team also developed an application to read various statistics on the network. It then makes a call to the Openflow controller to make changes the application deems necessary.
“The application we created is smart enough to dynamically make changes to the network based on data it receives from the network,” said Meath. “Sometimes, for example, an administrator may not be able to predict how much bandwidth is needed, but the network will know when to precisely add or remove bandwidth."
Also developed by the team is a web interface dubbed Avior. This enables a network administrator to add, delete or modify flows on the network—no matter where they are located, even on a mobile device.
Preventing network outages is the main focus of the invention.
“Let’s say that there is a virtual machine in a data centre in New York City that is streaming a sports event to TV viewers and suddenly a natural disaster, such as a hurricane or flood, is headed towards that data centre,” said Meath.
“With our invention, a network administrator could immediately and remotely migrate the virtual machine to another data centre in New Jersey, which is safe because it is outside the potential disaster area.”
An administrator can also opt to provide extra bandwidth to accommodate the virtual machine migration and re-provision optical links between the data centres to add bandwidth.
“Then, when the migration is complete, the software will automatically remove the unnecessary bandwidth. As all of this happens, a fan watching the game won’t even notice that anything has changed,” he said.
The invention is now being demonstrated to IBM’s customers and is expected to be commercially available in 2014.