We're not so different, you and I.
While Microsoft has pledged to let LinkedIn “retain its distinct brand, culture, and independence,” after its $26.2 billion acquisition of the biggest social network for professionals is closed, it is likely that its new parent will eventually want LinkedIn to adopt its data center strategy to match its own.
Microsoft transitioned to a uniform infrastructure strategy across all its product segments in the 2013-2014 time frame, going from a past model where every product group made its own IT decisions to the current one, where the company has several server SKUs optimized for various purposes, and its various groups have to choose from that list. The company switched to this approach because it was a more economical way to scale globally.
There are several recent examples where a company that’s been acquired eventually transitions its infrastructure strategy to match the parent company’s one. The process also usually includes consolidation of the data center resources, with the subsidiary moving its applications onto the parent’s infrastructure.
One example is Instagram, which in 2014 moved its applications from Amazon Web Services into Facebook’s data centers, about two years after Facebook acquired it. Another example is LinkedIn itself, which last year moved the application stack of SlideShare, the company it acquired several years prior, out of a managed hosting provider's data center into one of its own data centers.
Whether the process of LinkedIn’s absorption into Microsoft will include infrastructure consolidation is unclear at the moment. As the two examples above illustrate, companies can take several years to make such a decision and implement it. However, because of some changes LinkedIn recently made to its data center strategy, aligning with Microsoft’s will not be an end-to-end switch for LinkedIn’s infrastructure team.
LinkedIn has been transitioning to a hyperscale data center strategy, one that’s similar to the strategy used by Microsoft, Google, Facebook, and other companies that provide services over the internet at global scale. That strategy includes designing a lot of the technology that underpins its application in-house, including everything from custom data center cooling and power infrastructure to networking switches. It also includes shifting to a limited number of uniform server SKUs in all data centers.
- LinkedIn Switches to Custom Data Center Design
- LinkedIn Adopting Hyperscale Data Center Way
- LinkedIn Designs Own 100G Data Center Switch
The first facility where LinkedIn is applying this strategy is its new data center outside of Portland, which it leases from Infomart Data Centers. The second was a LinkedIn data center in Singapore, which came online in March, its first data center outside of the US, leased from Digital Realty Trust.
The company has been planning to transition its other data centers, located in California, Texas, and Virginia, to the hyperscale way in the future as well, Yuval Bachar, LinkedIn’s principal engineer of global infrastructure architecture, told us in an interview earlier.