IIIS: Melbourne IT's CTO lays out four-year roadmap

Melbourne IT may be able to offer discounts to some enterprise and SMB customers as backup/restoration requirements change in the future, according to the hosting provider's chief technology officer.

Speaking during the Implementing Information Infrastructure Symposium (IIIS) in Sydney this month — co-hosted by Computerworld Australia and SNIA ANZ — CTO Glenn Gore outlined the company's four year plan up to 2015. Melbourne IT's hosting business supports about 80,000 websites and generates 150 million Web page views per day.

Gore said that the company was looking to change its backup services. Currently every byte of information that is stored within Melbourne IT is backed up and restored but this is not sustainable going forward.

"From a price perspective, and where the industry is going with Cloud based hosting platforms, innovations like object-based storage will allow us to start considering different ways of protecting that information," Gore said.

"We can now tell customers that they are selective about what needs to be backed up they will get a great price differentiator in service."

For example, he said that the majority of its customers will much of their own information backed up internally.

"We still need to protect all customer data but we don't need to back up every single photo because they will have a copy stored on their own system," he said.

Gore also predicted that in 18 months, the company will be using object-based storage more often than block-based storage.

"What's great about object-based storage for us is that first of all it's a Web service driven approach and the only way we `talk' to the storage is by a Web API. It's simple and easy to work with."

"That's not saying we are going to replace block-based [storage]; one of the big challenges for an object-based environment is applications and most of the applications we use today do not know how to `talk' to object-based storage," Gore said.

"You are not going to be able to take your Oracle stack and just point them at your object-based storage."

Another big change for the company will be around disaster recovery; as people realise that the value of content is increasing rapidly and as the cost of storage drops, the value of that data is rapidly increasingly, according to Gore. "As the value of that data increases, people want better and more sophisticated ways of protecting that value across the storage platform. Using storage virtualization is replicating within the data centre. If I can send the data between the data centres really easily, it only solves half the problem because I have an inconsistent view of the data application perspective to multiple sites."

Gore said the Melbourne IT has gone through a cost reduction exercise, with a move to direct-attached storage which began in 2008.

"Back than, most of our services were on physical servers and virtualization had just started to come in at the server layer and we had this issue of physical data," he said. "Our data was stuck to a physical server and that caused us all types of issues, because if that server failed, as they quite often do, the data that was attached to that server was not retrievable until we got that server back up and running.

In addition to moving away from attached storage, Gore said the company wanted to increase the revenue of storage, because as a hosting provider every gigabyte of storage is directly linked to the revenue that it is producing.

"There is a great metric internally which is what percentage of our storage is producing revenue and what percentage is not attributable. So an example of that could be all the log files coming off the shared Web hosting farms, which includes a couple of terabytes a day which would, in the past, not have generated any revenue."

Melbourne IT's current environment has been consolidated from 30 storage area networks (SANs) down to nine.

Follow Hamish Barwick on Twitter: @HamishBarwick

Follow Computerworld Australia on Twitter: @ComputerworldAU

• Bookmark this page
• Share this article
  •  
  • facebook
  • slashdot
  • digg
  • Reddit
  • stumbleupon
  • linkedin
  • twitter
• Got more on this story? Email Computerworld
• Follow Computerworld on twitter

• Share
  Share this article

  •  google+
  • facebookfacebook
  • slashdotslashdot
  • diggdigg
  • RedditReddit
  • stumbleuponstumbleupon
  • linkedinlinkedin
  • twittertwitter
• print
• email
• Bookmark