Why do another startup? Why put in all the hours? Why go without pay for long periods?Frankly, we’re not done.
A decade after rolling out the first cloud platform (yes, AppLogic predated EC2), a lot of the changes we’d hoped cloud would enable still remain elusive. It’s become much easier to deal with provisioning. Few startups now raise money just to rent a cage and buy servers. APIs have put infrastructure in the hands of engineers. In addition, modern startups don’t have to add system administrators at the same rate as engineers. Actually, I’ve encountered many companies over the last few months with a single devops engineer responsible for all operations. Moreover, cloud API’s have made it commonplace to be able to start and stop entire services easily. So what’s missing? The answer to the question starts with a story.
At 3tera we designed our cloud platform, AppLogic, to allow users to allocate virtual machine resources in exacting detail. Memory could be assigned in 1MB increments, CPU was controlled to 1% of a core and bandwidth in 1Mbps. Our belief was that resource utilization would be important to customers. We were wrong. AWS offered a few predefined virtual machines sizes, and customers greatly preferred that simplicity. We learned two things from that experience. First, when the price got low enough efficiency was of no importance. Second, system administrators, developers and systems architects had no tools to use that would help them allocate resources in minute detail. Our users struggled because we offered tremendous flexibility without information about how to use that control.
We believe the information void we uncovered in the cloud era still exists today. That fact is exacerbated by the fact most infrastructure management still focuses on the system administrator, via a UI, as the central decision point. As companies move to DevOps, automating more and more of their operations, the combination results in uncertainty about how to deal with certain day to day decisions.
Datagrid was founded with the mission to provide actionable data for DevOps.