Disclaimer: I recently attended VMworld 2014 – SF. My flights and accommodation were paid for by myself, however VMware provided me with a free pass to the conference and various bits of swag. There is no requirement for me to blog about any of the content presented and I am not compensated in any way for my time at the event. Some materials presented were discussed under NDA and don’t form part of my blog posts, but could influence future discussions.
VAPP3031-SPO – Deploying Web-Scale Converged Infrastructure: Software Defined Storage for the Real World
VAPP3031-SPO was the last session today and, while it was sponsored by Nutanix, I found it reasonably useful to get a customer’s perspective on things, so I figured I’d at least write up my notes.
It was presented by:
- Sachin Chheda, Nutanix – Director of Product Marketing
- Ali Alladin, Williams-Sonoma – Global Director of Enterprise Infrastructure
- Dan Gorman, Williams-Sonoma – Senior Enterprise IT Architect
Williams-Sonoma are best known for the Pottery Barn range of stores, but they do a bit more than that. You can check them out here.
Nutanix, of course, are pretty well known. They were founded in 2009, and are now shipping the 4th version of their OS. Their stated goal is to “deliver IT infrastructure that delivers the scale of Google, Facebook, etc to enterprise IT”.
So what’s webscale?
- Scale one x86 server at a time
- Smart software, commodity hardware
- Highly distributed software
- Storage is local
- Heavily automated
So what is it not?
- Vanity hardware
- Islands of storage
- Unused, over-provisioned resources
- North-south network traffic
- Lots of server and storage admins
Nutanix are also focussed on what they call “VMCaliber Operations”
- Meet SLAs and provide cost-effective DR
- VM and infrastructure insight
- Non-disruptive operations
Sachin provided a brief explanation of the architecture, and then handed over to the Williams-Sonoma guys.
Their business is multi-channel, multi-brand – retail stores, catalogue, on-line. It’s therefore a complex IT infrastructure and they need to be reactive.
What drove the move to Nutanix?
- Moving to 100% virtualised infrastructure
- Converged infrastructure
- Deploy infrastructure for remote and branch offices
Looked at standalone, Open Compute Project, and FlexPod, but they needed a FRU model. They also had a lot of infrastructure silos.
They needed efficiency and ease of management, also needed replication. It also needed to
- Deliver performance
- Scale linearly
- Avoid noisy neighbours
For workloads such as:
- communications and collaboration
- Business intelligence
With Nutanix they got:
- Resiliency and availability
- Tiering and data localisation
- VM management and protection
The criteria was provide a consistent repeatable infrastructure design that was designed to survive failure.
So they did a PoC and it did what it said it would. Happy days.
They now have an IaaS model that their business units understand. They’re aiming to have an internal private cloud on Nutanix, and want to be able to scale in any direction as business changes.
One thing they did need to do was rethink management. The PRISM UI has plenty of pretty things to look at though. and they could also integrate REST API with tools such as PERL.
Some of the advantages realised included:
- Predictable performance
I liked this approach:
You can’t be efficient if you keep putting yourself in the loop. You need to start with intelligent IT infrastructure.
- Pay attention to your apps and the IO and data flow
- Build a reasonable level of competency
- Continue building out our service model
- Automate more functions
- Evaluate new workloads such as big data
- Find out your requirements – efficiency and scalability should be on your list
- Be thorough – investigate all of your options
- Remove yourself from the equation – rely on dashboards, automate where possible.
There were some nice customer insights here, without too much marketing fluff. 3.5 stars.