By John Katsaros
Sept. 14, 2015
If you ask ClusterHQ’s CEO Mark Davis how he will achieve his goals for world domination in the data center, he’d probably answer in 3 steps:
- #1 Build a really innovative product.
- #2 Create an active user base, and
- #3 Make no enemies, and partner with everyone.
Mark used a similar strategy as founder and CEO of Virsto, a software-defined storage start-up acquired by VMware in 2013: How to succeed in a VM industry dominated by VMware. Mark focused Virsto on making VMware’s VDI (virtual desktop infrastructure) more affordable.
For years VMware tried to crack open the VDI market, but the cost and complexity of the VMware solutions were too high — the deployment economics were insufficient for widespread adoption. Storage costs were the most expensive part of the overall VDI ROI model, and VDI storage costs were high because storage was hard to virtualize. So Mark aimed Virsto at reducing storage costs and cut the cost of storage per desktop by as much as 70%. That made it an easy target for a VMware acquisition.
Davis plans a similar but different approach to the container market. ClusterHQ’s focus is based on bringing basic container data management concepts to the container world – to enable production deployments of container-based systems. ClusterHQ’s open-source solution is called Flocker (download at clusterhq.com/flocker). Flocker enables containers and their associated data volumes to move between servers as a single unit, which is needed for Microservices.
This allows DevOps teams to relocate application containers as utilization conditions change so performance remains high. Flocker also increases the value of a container investment to a broader set of applications by making containers portable between applications.
Flocker lets DevOps teams run databases in Docker containers that can be as portable as application containers. Application and data containers can be moved between servers in a data center, between data centers or across private and hybrid clouds.
Flocker is container-agnostic – it will run all popular containers using gateways to accommodate each new platform partner. Flocker solves the problems that containers have had dealing with stateful data. The initial Flocker applications for containers are: moving a stateful container, live container migration, cross-datacenter migration, Availability Zone (AZ) failover, container back-up, end-to-end production testing, automatic failover, database cloning, high availability, disaster recovery and cloud arbitrage.
These mostly infrastructure-related applications are just the beginning of what DevOps need to efficiently run a container-based system. As Mark Davis might say, “There is no limit on the amount of software that needs to be created.” Flocker seems to be poised for rapid adoption as the container market evolves.