Blockdaemon = (Kubernetes + GCP)² + (a dash of secret sauce)

Konstantin Richter

Blockdaemon for Blockchain

Since we are the most operational DevOps shop in the blockchain space, we have unique insights when it comes to performance of decentralized networks and the respective abilities of different protocols.

We recently gave a talk at Google’s first Blockchain Event about our tech-stack evolution, and why we have moved our orchestration and monitoring over to Kubernetes/Terraform via GCP.

The more mainnets we are building for the leading projects needing to provide connector-end-points to enterprises (aka nodes), the more are we learning about how much is still missing for enterprises to be comfortable with network security.

To date, no enterprise would willingly submit internal data to functional external blockchain networks, besides data-siloed marketplaces that are run as experiments. The reason is that blockchain networks to date cannot provide a SLA or actually monitor nodes hosted across multiple different cloud providers and on-prem setups. Enterprises will only open up data when their security and compliance standards can be mirrored on decentralized networks. That creates a bunch of complex challenges.

Kubernetes for Blockchain

We are working with a few F500s who are actively seeking to move blockchain projects into production but cannot get sufficient insurance from providers about network performance. That is where Kubernetes helps, because all nodes running on it can be monitored in their respective containers. We are adding additional monitoring specific to nodes/blockchains via Prometheus and some other slick tools we found. For that we needed to learn what a healthy node looks like, and how you can “heal” them when they aren’t.

Luckily enough, working with Stellar, Algorand, Aion, Ethereum, Responsible Gold, GoChain asf. has taught as a lot — each protocol requires a different node configuration, but the generic health of a node at least falls into similar categories. So besides DevOps and monitoring of nodes run within the Blockdaemon cosmos, we will add monitoring for nodes that are run on-prem or non-Blockdaemon provided hosting, vastly expanding our market size.

While we love being the “Heroku” for blockchain, working with the Heroku Founders has taught us that we need to forge our own path, so we are adding features that move us towards the newrelicification of Blockdaemon. That said, direct brand comparisons don’t work for very long, unless you have no personae of yourself, so we will settle for being the Blockdaemon of decentralized infrastructure and let others be the Blockdaemon of Blockchain!

Decentralization works when monitored and dynamically secured.

Careful observers of our product and blogs will have noticed the decentralization score connected to network requirements that we launched last month. That was the first step towards bringing new-relic type features to blockchain via Kubernetes. The monitoring via Prometheus and an ability to “heal” nodes is nearly complete. Following that, BYOI will go live via our API/client. What excites us about all this is its potential to give meaning to decentralized networks, because our auto-scaling features will allow us to dynamically ensure a % decentralization for permissioned networks. This includes attaching a node to non-blockchain networks for independent monitoring and observation.

Our Kinetic-Ledger sits on top of all that, so will add an additional simple abstraction layer for professional data use-cases requiring a specific audit-function. Simple but secure with trust embedded and guaranteed by a SLA. The on-ramp to shared databases has never been easier.

Nodes are becoming more popular

We have seen enormous traction of our service over the last 90 days — it took us 12 months to get us to 500 users, and 15 months to get to 5000 (so 4500 over the last 90 days), in addition to working with some of the most respective enterprise grade projects in the space like Algorand and SpringLabs.

We will have to automate the on-boarding of new protocols and find new ways to incorporate updates and fixes, empowering the community to help us ensure the best performance grade for nodes. We simply can’t respond to the many requests we are getting to support new configurations and protocols. The increase in usage corresponds with Crypto-Winter, and interesting correlations we talked about this week at Bloomberg LA. There are still plenty of well funded projects (anyone raising funds prior to Aug 2017 is still in the plus), and they are under more pressure than before to go to mainnet. The revolution isn’t quite dead yet.