Twitter will open its own data center

Recently, Twitter users have observed that the social network could not cope with the workload and experienced reliability problems. This was partly due to the holding of the World Cup in South Africa, partly due to the rapid growth in recent times - more than 300,000 new accounts have recently been created. This pushed the company to create its own data warehouse in Salt Lake City, USA.

Undoubtedly, the future data center will not be as big as Apple’s data processing and storage center in North Carolina. The project for Twitter was developed to order, and the launch of the center is scheduled before the end of this year.

Representatives of the company believe that their own data center will provide an opportunity to better and faster adapt to the growth in the number and activity of service users.

Twitter will have full control over the network and the configuration of systems that occupy a much larger area than now in a building specially designed for these capacities and cooling systems. The data center is assembled from equipment from different manufacturers and will operate under the control of free OS and applications.
')
Until now, Twitter has used data processing and storage centers built by NTT America in San Francisco Bay. The company will continue to work with NTT, which will manage the areas occupied by the equipment of a specialized data center.

This step followed Facebook’s January announcement to create its own data center in Oregon, where many companies have data processing and storage centers, including Amazon and Google. The reasons for choosing such a geographical location are cheap electricity, a good climate, and tax breaks for companies building data centers in this state.

Now Twitter is under massive public pressure, due to the need to explain to users why the service continues to fail and why some features are disabled. The company even posted a post in its own blog, telling in general about the problems and another post for technical specialists, describing the situation in detail. The main point of the publications is that on Monday one of the main databases was “stuck” on one of the queries and the system was switched to the limited capacity mode. I had to restart the database server, which took more than 12 hours.

It seems the company really needs more control over its own systems. And now it is clear why the company often compares the tasks of scaling, maintaining and fine-tuning its service with the work on the creation of a medium-range missile.

According to TechCrunch