AOL has released details of a new unmanned datacentre, called ATC, which is able to deploy virtual machines in less than eight seconds with advanced automation technology.
ATC is currently being used by AOL for its internal services, but could this mark the beginning of automated commercial datacentre services provided by the big cloud providers, for example?
The datacentre took just 90 days to build using open source code, and Michael Manos, VP of technology operations at AOL, argues that it "represents a model on how to migrate the old [and] prepare for the new".
How does it work?
In a recent blog posting, Manos states that the company was able to operate a 100 per cent lights out facility (no staff on site) owing to a change in planning.
"There are absolutely no employees stationed at the facility full time, contract, or otherwise. At other facilities, there are always personnel on site," he said.
AOL is using its own advanced configuration management (ACM) system to monitor the datacentre, and this helps with automation of processing resources.
Although all organisations will invest in ACM, AOL has taken the traditional architecture a stage further to enable an unmanned datacentre.
"All organisations invest significantly into this type of platform.
If implemented correctly it can be used to provide more than rudimentary asset management, to include cost allocation, dependency mapping, detailed configuration and environmental data," he said.
Manos indicates that the speed at which virtual machines can be deployed in ATC is down to the configuration management system it has developed.
"We went from provisioning servers in days to getting base virtual machines up and running in less than eight seconds," he said.
AOL's ability to remain staff-free at ATC was put to the test when a 5.8 magnitude earthquake hit the East Coast of the US and the AOL homepage saw an incredible surge in demand.
Manos suggests that in the past such an event would have required a "massive people effort to provision more capacity", but thanks to the configuration management system, AOL was able to start adding additional machines immediately.
Would it suit cloud service providers?
It appears AOL has developed a fairly resilient, effective solution for operating an internal datacentre. However, if you were an enterprise using infrastructure services from a public cloud vendor, would a staff-free datacentre be a suitable solution?
Ian Brown, analyst at Ovum, argues that this type of system will never make it fully into the commercial space and better suits internal datacentres.
"Most commercial datacentres will need on-site staff to enable customers to deploy systems and connect up to networks, both WAN and LAN. They will also have on-site security, as there may be multiple customers who need to access the facility," said Brown.
"They will also have personnel on site to provide 'remote hands' support for customers in order to re-boot systems, remove and replace failed equipment and physically install devices into the racks."
Brown said that remote automation is unusual for datacentre facilities and infrastructure, but that technologies are being developed in this area and that there are certain functionalities to which it is more suited.
"Power and cooling is becoming increasingly automated, especially with the advent of free or ambient air cooling, which monitors temperature, cooling, air flow and humidity, where this is all monitored constantly and adjustments are made automatically," he said.
"There are also physical security systems in datacentres, which again are digitised and can be monitored remotely."
Brown believes that big cloud providers such as Amazon will see the "trickle" of automation technologies into their datacentres. However, they will never be staff-free.
"All commercial datacentres will retain a skeleton staff for maintenance, network services, system deployment and security," he said.
"There are things that AOL may not have to do for a datacentre it owns and runs by itself, on its own campus."
By eliminating high entry costs for big data analysis, you can convert more raw data into valuable business insight.
A discussion of the "risk perception gap", its implications and how it can be closed