I’m wrapping up 8 weeks on the road with customers (more than 10 states, Australia, New Zealand, Mexico, France, Germany, Italy, UK) – a ton of amazing conversations, listening and learning.
While people/cultures/languages/technology adoption differ – some very clear trends:
- Every customer is running (not walking!) to use All-Flash Arrays for transactional block workloads, particularly those that need consistently low latency, and are dedupe/compression friendly. The XtremIO business for us is absolutely on fire – and I’m seeing this now globally. I’m particularly interested to see how quickly the “new technology” PoC demand (a couple quarters ago, it was nearly 100% of cases had a PoC – now it’s a fraction of the cases) and direct vs. partner (in Q1, every case was “high touch direct” – now, many partners are simply doing XtremIO themselves). All really good things!
- Every customer is looking at “software storage stacks + commodity hardware” for broad swaths of workloads that have been traditionally supported by hardware appliances. Customer reaction to ScaleIO and the ViPR object stack is very positive. Feedback is generally that they are technical better than existing opensource alternatives (think Ceph as an example), but that EMC needs to get better at getting the software out to the community in easily accessible, routes for free (and perhaps even opensource) ways. I’m listening – and expect to see progress here!
- Everyone is getting that “the answer to infrastructure” is get to IaaS as fast as possible (including hybrid, APIs, service catalogs).
- Everyone is getting that that is about simplifying infrastructure (using converged/hyperconverged approaches) + Management and Orchestration (vCloud Suite/Openstack) + building real “InfraOps” skill into their ops teams (my name for the Ops side of DevOps).
- Seeing more and more Openstack in all places. We’re all over it (broad cinder integration including SDS via ScaleIO, contributions around Cinder, Manila, Nova, Neutron – and a Swift implementation in the ViPR object stack) – but can go even faster, and we will!
- People are also “getting” that Openstack is not well-placed when forced into supporting “classic Platform 2” application stacks that have deeper infrastructure resilience requirements – but rather leaning towards “Platform 3” app stacks with less infra resilience. I think people are starting to get that it’s not “VMware vs. Openstack”, but really an “and” answer vs. an “or” answer. I’m using the fact that clustering support across Nova instances is so primitive to make this clear fast. People will have MSCS in use in enterprises for a LOOONG time. vSphere/vCloud suite support for that is strong. If you’re building a new app – you don’t build clustering into infrastructure – you build app-level and data-fabric level resilience, and then things like Openstack and Docker become much, much more useful. It’s an “AND” people, not an “OR”!
- More and more Converged infrastructure in all forms. This isn’t about ONE WAY, but rather a “portfolio”: Integrated Infrastructure (think Vblock style designs), SDDC + BYO commodity hardware (think VMware SDDC, Openstack SDDC), hyper-converged appliances that are “common modular building blocks” (think Mystic, Nutanix, Simplivity) that “start small, hyper simple, etc”, and Rack Scale Infrastructures (these like common modular building blocks use software storage stacks that are hyper-converged, but have more hardware variation, as well as abstraction variation). Lots on this front, expect a blog post or two soon!
- Early stage stuff thinking is emerging around “how do I broker/cloak across multiple public cloud and internal private cloud APIs and capabilities?”
- Everyone is starting to think/work beyond IaaS to PaaS with things like Cloud Foundry.
- Everyone is starting to actually move forward on Analytics and Big/Fast data models and data lakes. Interesting early movement around Spark.
From around the world, a small set of neat pictures that are about use of new technologies:
At a customer loving XtremIO!:
(pssst – there is a new dedupe/compression Mitrend assessment tool now, useful when looking at XtremIO, more here)
The SAP team as a customer celebrating moving ViPR into production (love the words):
The first ever “official” DSSD unit shipping to a customer (we are still a ways out from our formal directed availability period) – this team, and this technology rocks!
The first ever ECS unit being shipped in July:
… to the first customer, the Vatican Library (who I visited in Italy – what a cool customer, and a great partner in Luciano Ammenti who is the CIO there) – here’s me with the GREAT SE (Lorenzo Puopolo) who supports the Vatican, and with Marco Fanizzi (great sales leader in Italy) and Barbara Jatta (with the Vatican Library) :
Technology is key – but in the end, it’s all about customers! Home for a few days, and then off to VMworld – working furiously on the content for what will be an epic VMworld, and there are more than a few surprises in store!
Great XtremIO photo, even though you look a bit stunned by the glow... Also, the next time we meet, please allow me to teach you how to properly knot a tie. A world traveler, like yourself, should always look your best! :)
Posted by: Alexander BREViiARIO | August 14, 2014 at 05:39 PM
I can't wait to see that DSSD box in action!
Posted by: Derrel Isler | August 15, 2014 at 09:50 AM