With each passing year, fresh stories arrive about major consumer internet companies and cloud app developers deciding that they would rather have a cloud giant handle much or all of their infrastructure needs.
The latest such story involves Twitter (TWTR) - Get Report, which announced this morning that it will be using Amazon Web Services (AWS) to serve up Twitter timelines to users, through a new multi-year deal. Amazon.com (AMZN) - Get Report indicates Twitter will be using a variety of AWS offerings, including computing, storage, app container and security services.
Notably, Twitter’s use of AWS computing services will extend to computing instances powered by Amazon’s proprietary, Arm-architecture, Graviton2 CPUs (offered as a low-cost alternative to instances powered by Intel and AMD CPUs).
As Amazon notes, Twitter isn’t a stranger to AWS: The company has already been using offerings such as AWS’ CloudFront content delivery network (CDN) and its DynamoDB database, which competes against MongoDB’s database. In recent years, Twitter has also been using Google’s (GOOG) - Get Report public cloud services to handle certain big data/analytics workloads.
However, Twitter has also maintained a large internal data center infrastructure to help deliver its services to hundreds of millions of consumers. Migrating what’s in many ways its core service to a third-party cloud signals a major strategy change, even if it doesn’t mean that all of Twitter’s data centers will be shuttered.
And Twitter is hardly alone among well-known Internet and cloud firms in making such a move. In recent years, cloud app developers such as Salesforce.com (CRM) - Get Report, Workday (WDAY) - Get Report and Atlassian (TEAM) - Get Report have also begun leaning heavily on AWS for their infrastructure needs, while Adobe (ADBE) - Get Report now relies heavily on both AWS and Microsoft Azure (MSFT) - Get Report. Spotify (SPOT) - Get Report, meanwhile, has migrated most of its services to Google’s cloud.
There are also numerous internet/cloud companies that have been relying heavily on public clouds from the very beginning, or close to it. This list includes names such as Snap (SNAP) - Get Report, Pinterest (PINS) - Get Report, Box (BOX) - Get Report and Snowflake (SNOW) - Get Report.
Netflix (NFLX) - Get Report also qualifies to an extent: The streaming giant runs its own CDN, but otherwise relies heavily on AWS to keep its services running, while also using Google’s cloud services to a lesser degree.
Some exceptions to the rule do exist. Dropbox (DBX) - Get Report, for example, migrated its cloud storage services from AWS to its own infrastructure a few years ago. However, as a Wired article from 2016 spells out, this effort required a considerable amount of heavy-duty engineering work.
By contrast, Twitter, Atlassian, Spotify and many other internet/cloud firms have decided that it’s in their interests to go in the opposite direction. While the exact motivations for a cloud migration vary from company to company, some of the common ones tend to be a desire to use various cutting-edge services available on public clouds, a desire to leverage the geographic reach of major public clouds to better serve international users and a wish to free up the time that a company’s execs, developers and IT admins have been devoting to managing an internal data center infrastructure.
And with Amazon, Microsoft and Google’s public clouds each steadily expanding in terms of both their scale and the number of advanced services that they offer, those selling points definitely aren’t getting any weaker.