LinkedIn’s multi-year migration deal with Microsoft Azure

LinkedIn, this week, announced its multi-year migration strategy on Microsoft Azure. This strategy was seen after much consideration by the social media giant towards bringing in the next evolution to its infrastructure. Let’s bring some facts and figures on the table and how the transformation phase will look like, considering that this move is the most prominent digital transformation for the company.

Ten years ago, LinkedIn had about 50 million members, and in the subsequent five years, that number jumped to 300 million. Today the business social network has more than 645 million members, 20 million jobs, and fun fact — someone is hired using LinkedIn every eight seconds. The scalability for the platform to handle massive data, stream video, and share media, can only be achieved through the public cloud. As both the companies are owned by Microsoft, so the decision seems natural.

LinkedIn has maintained its own data centre infrastructure even following its acquisition by Microsoft in 2016. In a recent blog post, Mohak Shroff (Senior VP of Engineering at LinkedIn) said the transformation might take multiple years to complete.
He further wrote, “moving to Azure will give us access to a wide array of hardware and software innovations, and unprecedented global scale. This transformation will position us to focus on areas where we can deliver unique values to our members and customers. The cloud holds the future for us, and we are confident that Azure is the right platform to build on for years to come.”

The decision comes after using several Azure services for the past few years. The services made a notable impact on LinkedIn’s business, which includes:

1. Azure Media Services:

Videos have become an integral part of LinkedIn in terms of content consumption. This service has provided LinkedIn with the agility, capacity, and elasticity required to accelerate LinkedIn’s video post-delivery.

2. Microsoft Text Analytics API:

More than half of Linked In’s active members live outside of the US, which underlines the challenges due to language barriers. So, they leveraged Microsoft Text Analytics API, an Azure Cognitive Service that can detect up to 120 languages, to improve machine translation in the feed.

3. Content Moderator:

LinkedIn tries to keep the feed relevant by identifying unprofessional and spammy content. To achieve this, they have established a two-way integration between Content Moderator, a Microsoft Cognitive Service running on Azure, and Linked In’s Universal Content Filtering (UCF) platform.

Azure benefits that LinkedIn finds tempting

Linkedin has four leased colo (co-location) space datacentres, with three of them located in the US (Texas, Virginia, and Oregon), and one is in Singapore. The primary benefit for Linkedin with this deal is the growing number of regions becoming available to LinkedIn. This move will make www.linkedin.com hosted on Azure infrastructure function at very less or no downtime. It offers more accessibility to build a more diverse and global platform.

The elasticity by the public cloud is another benefit that LinkedIn was looking from the corner of the eye, in terms of balancing the peak workloads by provisioning infrastructure requirements on the pay-as-you-go model.

Last but not least, the most significant benefit that Azure has to offer is the ability to focus more on building the product and enhancing the security of the site.

LinkedIn never promised any of the cloud partnerships at the time of acquisition. The use of services over-time, ever-increasing workloads, and the fear of bearing data centre expenses collectively directed towards primary use of the public cloud. The enormous investments for the digital transformation once again highlight the bright future of sophisticated technology.