Containers & VMs: Life as a winning team
Approximately a year ago a number of VMware blogs proclaimed victory – Docker and VMware are better together! The question is – is this a true statement? Is there complete synergy between the two – or is each looking to trump the other to ensure dominance in the enterprise market?
After watching an extensive overview from Nicholas Gerasimatos from FICO at Gartner Catalyst who orchestrates VMware and containers with OpenStack, my view is that there is ample home for each and BOTH are valuable together for enterprise compute environments. (Interesting data points – 90% of lending decisions in the US rely on FICO scores – and – FICO is at its core a big data company). On the eve of VMworld, I’m anxious and excited to hear more about how the last 12 months have played out in real customer environments. Will we see more proof points as to how a virtualized environment and containerized applications will transform how we roll out infrastructure and applications?
VMs and containers are easily distinguishable from each other. There are ample depictions available on the internet such as this one that indicates how each operates:
Containers are focused at the application level more so than infrastructure. VMware might argue differently, but I think it is easy to give app-advantage to containers simply by understanding the granularity at which they can operate. Microservices architectures – those in which components of an app can be broken down, ran, and scaled independently are well served by container technology whereas this is not in the wheelhouse of vSphere. Clearly the abstraction and efficient management of underlying hardware components – CPU, memory, etc. – are still the strong suit of VMware vSphere virtualization. As it pertains to abstracting hardware, who can argue that VMware is king? But, when it comes to packaging and distributing APPS it looks clear that containers are the right approach.
When all is said and done – the other significant point I take from Nick’s talk is that storage was no walk in the park. FICO thought they could get it done with CEPH – then had to adopt a completely different solution when performance became important. What did this mean to FICO? They got what they needed, but they had to do it at a cost – the cost of managing multiple storage solutions. They though they could do it with one, but couldn’t.
The story need not end here. This is where Hedvig has set its sights. We know that storage of different performance levels, different protocols, and different characteristics are needed for app X vs. app Y – but can we get this all from one solution? Industry pundits will say NO – but we say YES. Hedvig’s day-zero design principles were based on the understanding that users would need:
- Storage of differing performance levels – varying IOPS delivered by RAM, flash, HDD, etc.
- Storage of differing protocols – block (iSCSI), file (NFS, SMB), and object (S3, Swift)
- Storage with enterprise features – server-side caching, dedupe, compression, replication, DR
- Storage that’s easy to setup and manage – vCenter integration, point and click provisioning, minutes-to-ready
We aim to please. Come check out Hedvig to see how we fit the equation when it comes to serving virtualization, container, and bare-metal environments. This week at VMworld we join ClusterHQ to discuss virtualization and containers. ClusterHQ with Flocker enables storage portability with container environments and Hedvig delivers software-defined storage with maximum flexibility, elasticity, and simplicity.
We want to meet you! You can visit us this week and see a technology demonstration at ClusterHQ booth 2636 at:
- Monday, August 31 from 1:30 – 1:45 pm
- Tuesday, September 1 from 5:30 – 5:45 pm
- Wednesday, September 2 from 3:30 – 3:45 pm.