Atlassian's Stash Data Center Brings Enterprise-Grade Git to Compute Clusters

Atlassian’s Git repository management tool Stash has destroyed scaling limitations. The new enterprise offering is called Stash Data Center, and it can now scale to tremendous heights using active-active clustering.

The large Australian software company is best known for its popular project management software JIRA, but it has a number of other products and services, including Stash, the enterprise-grade distribution of Git, an open source system for code revision control and management.

There’s a new culture overtaking even the largest enterprises, general manager of Atlassian’s developer tools business unit Eric Wittman said. Development is adopting a more agile methodology in general and collaborating across larger teams. It is development with real-time communications.

Stash is Atlassian's on-premises source code management for Git. In a secure and fast setup, customers can create and manage repositories, set up fine-grained permissions and collaborate on code.

Stash Data Centers is Stash that works across server clusters that can be seamlessly scaled, unbeknownst to the developer. A lot of automation is happening around source code, which in general means servers are under a ton of duress, and Stash Data Centers' clustering capabilities eliminate this problem.

“Stash Data Center is the enterprise flavor,” said Wittman. “It can run on a cluster instead of a single server, easily supporting 10,000 developers. If I want to add an additional node, I can do that without taking the whole system down. This scaling issue is a very complicated problem, and we are the first to solve it.”

While Stash was already fast and secure, Stash Data Center removes scaling limitations. This makes for a promising offering to enterprises looking to establish agile development workflows. Each node in the cluster increases capacity for concurrent users without sacrificing performance.

The product comes with enterprise-grade security and collaborative workflows. Repositories are protected by the user's firewalls and permissions are customizable at the global, project, repository and branch levels.

It lets customers use any form of load-balancing tech, be it software or hardware, to intelligently distribute the load. Data center deployments integrate with industry-standard technologies for database clustering and shared files systems to minimize single points of failure.

It adds application resilience: users can increase application throughput to avoid performance degradation in the event of load spikes. Daily workflows don’t change for users, but they’ll see  fewer slowdowns, faster compile times and less downtime.

Atlassian started 12 years ago and spread primarily through word of mouth. It doesn’t spend a lot on marketing and a sizable portion of the company’s money goes into R&D (about 40 percent, according to Wittman). “It’s a large percentage for a large company,” said Wittman. “We basically allocated almost half the team to developing Stash Data Center.”

The company spent a lot of R&D removing the scaling limitation. Atlassian already touts more than 40,000 organizations using its products across a multitude of industries. And the numbers aren’t fudged: if three departments within Microsoft are using Atlassian, it’s counted as one customer.