Don't look now, but there's a daisy chain being built in the cloud.
In January, Verizon bought the pioneer cloud company Terremark for $1.4 billion. Three months later, CenturyLink announced it would buy managed services host (and cloud infrastructure provider) Savvis for $2.5 billion. In June, a subsidiary of Japan's telecom giant, NTT, bought Opsource to establish a cloud solutions business unit.
What do you get when you marry cloud data centers to a telecommunications company? My answer is: the beginnings of a cloud network, a chain of linked data centers that in some cases bring two or more data centers into a position of backing each other up. That's something many enterprises would value as they move workloads into the cloud. The service freeze in Amazon Web Services' northern Virginia center over the Easter weekend served as a reminder of the value of geographic distribution when it comes to backup and recovery.
I hope Amazon Web Services is studying the trend. If you were an AWS customer in US East-1 (northern Va.), you couldn't easily designate the AWS data center in Dublin as your preferred failover site. You couldn't even select AWS' US West data center, unless you constructed the network links to it yourself. You were stuck using a neighboring Amazon "availability zone," which, it turns out April 22-24, was not necessarily in a separate data center and in some cases froze up at the same time as your primary zone did.
Automated backup to a separate location, however, is a definite possibility, given linked chains in the cloud. The CenturyLink-Savvis combination brought together 34 Savvis and 16 CenturyLink data centers in North America, Europe, and Asia. The Verizon-Terremark union combined 13 Terremark data centers with 36 Verizon hosted services data centers, for a chain of 49 in North America, South America, Europe, and Asia.
These are not literally daisy chains, with one data center tightly linked to the next in an orderly progression. Rather, they are regionalized groups of data centers, with one facility serving as a hub location. The hub acts as a central facility for connecting all members of the group to the Internet and other network junctions via high speed pipes.
Geographically separated data centers have been linked before, of course, but the new combinations are being interlinked by a single telecommunications company that can offer automatically implemented connections as a service over high speed lines.
I asked Ben Stewart, senior VP of facilities engineering at Terremark, how one of its customers would have used a distant Verizon data center as a backup site prior to the merger. It could be done, he said, but it would have taken a lot of work by a telecom-savvy cloud customer. First, he'd have to study the rate tables of various telecom carriers that were available on the route, then negotiate with each for network segments that could serve his need, then test it to see if it all worked.
If something went wrong--such as a router was failing at a key network junction, generating thousands of retries--he might have to make a half dozen phone calls to find out where the trouble was that was slowing his operations.
Now if a Terremark customer wants to link to another data center, "it's a click on a box of an order form," and the link and additional account will be set up. Furthermore, the number of carriers involved and the number of router hops has been drastically reduced, improving speed.
Terremark had initiated this approach on its own with its 13 data centers, but "it would have taken us a long time to achieve global reach," Stewart says. "Now, with Verizon, we're there."
For example, Terremark on Sept. 22 opened a new 25,000 square-foot data center in Amsterdam, where it has quick access over a 20-Gbps line to one of the world's largest Internet exchanges. The new facility serves as a hub for four other Terremark data centers in the region and provides a one-hop connection for them to the Internet.
How does the Amsterdam network access point change the picture for Terremark's other data centers around Europe, where it already has a strong presence? It brings London, Frankfurt, and Madrid closer to many other parts of the world.
Amsterdam is directly connected to Terremark's Network Access Point of the Americas, a huge, 750,000 square-foot facility in downtown Miami. So much network traffic converges on the network loops that circle the downtown that Miami is one of the five best interconnected cities in the world, Terremark says on its Web site. It's a gateway to other points in the U.S.; from NAP of the Americas, you are one router hop from Bogota, Columbia, or in two hops from Sao Paulo, Brazil.
"Amsterdam to Miami is one hop at two-thirds the speed of light over fiber cable. So Terremark customers in London, Madrid, Paris and Frankfurt are all linked now, just two hops apart through Amsterdam. They're just 3 hops away from Bogota," says Stewart.
Global companies, such as InterContinental Hotels, are looking to decentralize their mainframe applications and make them available to serve customers in many different locations. Bryson Koehler, senior VP of revenue and guest information, knows that if InterContinental's booking and customer relationship systems can be distributed, service to customers will seem instantaneous and give his firm a competitive edge. He literally wants the customer to be able to check into his room from a smart phone while riding in the taxi from the airport. To do that, he must find a way to offer distributed services. He's working on it, but not ready to distribute.
Running coordinated systems in a distributed fashion is difficult to do today, but infrastructure as a service in the form of interconnected data centers could make it a check box option. Having established a workload in the cloud, just check off where else you would like to have it run, and indicate which one is to serve as the primary copy.
Daisy-chained data centers in the cloud will one day host carbon copies of key enterprise applications. In some competitive situations, including the hotel business, distributed applications will make a competitive difference.
Verizon-Terremark is my primary example in part because, in addition to the 49 data centers that make up the Terremark unit's infrastructure as a service, Verizon also operates 200 telecommunications data centers that handle billing and customer service for its phone service customers. A telecom data center isn't much different from a cloud data center. In the future, it's possible Verizon or other telecoms will reserve space in their existing data centers to give their data center chain more locations for distributed systems.
CenturyLink-Savvis is another IaaS and telecom supplier and that combination has 50 data centers around the world, another example of interlinked data centers with access to key network junctions and exchanges. Amazon builds its data centers at prime network junctions as well, but at this point, it's still up to the customer to navigate between one center and another.
In the future, no cloud data center will be an island. No tsunami, hurricane, or earthquake will be able to take down your systems, just because the primary data center in which they were located lost its power or had its operations disrupted.
Both virtualized systems and cloud workloads can migrate quickly to a new location. A running virtual machine can transfer to a new home in milliseconds; given a few seconds warning, it can migrate with no data lost. Virtualized systems running inside a data center that is part of a chain in the cloud may prove to be a much more durable and available system than some of its predecessor high availability designs.
Most companies have little control over their heaps of unstructured data, our research reveals. It's time for a content management strategy. Get the new, all-digital InformationWeek supplement. Download it now. (Free with registration.)