Hello Steemians, we are excited to announce that Steemit’s APIs are now powered by Hivemind! By routing most of the social APIs to Hivemind, we are able to dramatically shrink the size of the full nodes we run. Because there are so many social applications sharing information on Steem, full nodes had become extremely large and expensive to run especially because they are not optimized for use cases that are not consensus-critical.

Hivemind is a service that syncs a traditional database with the blockchain. An added benefit of Hivemind is that since it’s written in Python, developers can modify its behavior and APIs with ease, instead of relying on C++ developers and modifying code which can interact with critical consensus logic. Implementing features like unreblogging becomes trivial with Hivemind while simultaneously enabling node operators to dramatically shrink the size (and cost) of their servers.

Testing

In our tests we were able to route all api.steem.com APIs to the new configuration—Hivemind for social APIs and everything else to “slimmed down” steemd nodes—which performed well. In addition, these nodes were running on-disk, as opposed to RAM; a major win which was only made possible thanks to AppBase.

We are extremely excited to see Hivemind running in production, on-disk, and utilizing AppBase. After completing this task, we have decreased our steemd instance sizes from 488GB RAM to 61GB RAM instances. We are doing further analysis and may be able to decrease this usage even further. These are major improvements to Steem that will make it easier and more cost efficient for anyone to run a full node, and all without requiring a hardfork.

Bug Hunting

We have so far identified one minor stability issue and will be rolling out an update shortly. If you are experiencing any issues using Steemit.com please leave a note in the comments.

The Steemit Team