Disclaimer: I recently attended the EMC Forum 2013 – Sydney. My flights and accommodation were paid for by myself, however EMC provided meals and some 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.
Rather than give you an edited transcript of the sessions I attended, I thought it would be easier if I pointed out some of the highlights. I hope to do some more detailed posts in the near future. This is the first time I’ve been to an EMC Forum since I started my blog, so it was a bit different to listen to what was going on in terms of things that would be blog-worthy. If it comes across as a bit of propaganda from EMC, well, it was their show. There was some really good information presented on the day and I don’t think I could do it justice in one post. And the ViPR logo is still one of the coolest logos for tech stuff that I’ve seen in a while. Note also that there was a metric shit-tonne of buzzwords used in this keynote, so bear with me. That’s an observation, not a criticism. I know the keynote has to appeal to both techs and suits (and those in between). Still, it was a lot.
Shaun McLagan – General Manager, RSA ANZ – was the MC for the keynote. He started by talking about the fact that the EMC Forum has been running in ANZ for the last 10 years. EMC shares have risen 281% in that time. There’s also been lots of change in that time. We should think about the power of change and what that means for us.
Alister Dias – VP and MD, EMC ANZ started off by talking about four key trends: mobile; cloud; big data; and social. EMC is apparently laser-focused on cloud and big data. These four trends underpin the shift to what they’re calling the “third platform”. The first platform being the mainframe, the second being client-server. The third platform is everywhere, and is driven by mobile. New order of magnitude with billions of users and hundreds of millions of applications. Pivotal was launched to take advantage of the third platform. EMC is growing, this is their 15th consecutive quarter of growth. Customers are focused on driving efficiency and reducing operational costs. IT is still too heavily oriented towards maintenance, not innovation.
Not all workloads are equal. This variety is driving EMC’s infrastructure innovation. Customers are starting to use private clouds, virtual private hosted and public cloud solutions. ANZ are early adopters, and thus more aggressive in shifting to these models. EMC is trying to leverage the software-defined (?) network and storage stack in the same way as VMware did with server virtualisation / compute. ViPR is EMC’s software play that provides the ability to abstract, pool and automate storage. Not just EMC storage, anyone’s storage, even commodity storage. Universal remote control (!) to provide the ability to manage this storage from one point.
In the year 2000 we created 2000PB of information during the year. Today, we create 4000PB of information in a single day. This apparently where big data tools help us harness the information to provide better service and so forth. Rapid deployment of apps, accessed anywhere and at any time. New apps will need to not only leverage structured information, but also unstructured data. Telemetric sensors are also going to be useful sources of data. Cisco has been counting internet connections. At the start of the year it was around 8 billion. By the middle of July the number had risen to 10 billion. The internet of things feeding information back to the machine overlords (I may have made that bit up). Apparently, by the year 2020, 40% of all information generated will be from these sensors.
Pivotal is fundamentally designed to provide customers with the tools to take advantage of the third platform. GE bought 10% of the company, because they’re into that sort of thing. movideo apparently uses EMC’s big data platforms to analyse metrics of viewer details in real-time.
Cloud and big data are only going to work if EMC provides a “blanket of trust” to organisations, via RSA of course. With integrated backup and recovery and global high availability. Now is the time to separate the hype from reality *cough*.
Shaun then takes the stage to introduce John Roese – Senior VP and CTO, EMC. John starts off with a bit of his career history and why he’s a fan of the ANZ market (early adopters who exploit technology to its full extent). When EMC approached him, he said he wasn’t a storage guy. That’s okay, EMC’s in the information management business, the big data business, the virtualised infrastructure business, the security business, and so on.
What’s the purpose of infrastructure? To empower workloads (the applications, the data, the process). The problem with workloads is that they change and evolve. Second platform application growth will be about 70% between 2012 and 2016. A new class – social, mobile, collaborative – is emerging. In the same timeframe we’ll be looking at 700% growth in third platform workloads. Each workload also has a different expectation of infrastructure. If we build purpose-built infrastructure for every application, things can become complex. Third platform doesn’t replace second platform, although some workloads may move to third platform environments. Second platform IT builds infrastructure to scale for the number of employees accessing it. Third platform builds to scale to the number of customers. If we accept that the third platform is new, then we need to accept that appropriate infrastructure for this platform will be new as well. Converged and software-defined whatever will provide us with new tools to meet these requirements. We need to think differently about how we do things. We can’t backup an Exabyte of data across a network. Instead, we need to look at what we do with that data in the first place, how we create it in a resilient fashion. John then spent a little time using healthcare as an example of how the use of data and infrastructure are changing.
There are three things we can do to keep up with changing IT. Firstly, stop thinking of infrastructure as a collection of boxes. Think of infrastructure in the future as something that can deliver pools of capacity and pools of performance-optimised infrastructure. Every workload needs one or the other or both. Secondly, because infrastructure is complex, we need to be working on ways to abstract them and simplify it via software (like VMware did). Thirdly, IT can use the information it has available to provide value to the business, thus adding value to IT itself.
Today, when we process data, we move it to the application and the compute and let it do the processing. But wouldn’t it be simpler, when you have a PB or EB of data, to move the application to the data to do the processing? Apparently there wasn’t really a performance tier until recently, as the compute was the bottleneck, and the application demand wasn’t there. Is generally low-latency, high compute / storage affinity or locality, scales to TB, not PB (costs a lot), and is biased towards flash. In the future, you’ll have capacity and performance tiers in on-premise and off-premise configurations. The complexity comes from how well this all works together. Chris Evans did a nice article on EMC’s potential misunderstanding of what’s really going on over at his blog, while Martin Glassborow also did a good piece on his blog from a slightly different angle.
EMC have $4 billion a year to spend on R & D and M & A. They say they will lead in performance tier innovation. First to put flash in an array, and nowadays you can buy a full flash VNX. They have XtremSF and XtremIO. They also bought ScaleIO a few weeks ago. As far as EMC are concerned they are number one in the capacity tier. The key is going to be staying number one. Mentioned a major refresh to VNX is coming in a few weeks.
With all of this capability comes a large amount of complexity. This is where something like ViPR comes in – the software-defined storage solution. Storage becomes less about plumbing, and more about delivering a very specific experience for each of your application workloads. The industry got storage virtualisation wrong the first couple of times by putting a heavyweight software layer in between perfectly good storage and perfectly good applications. The virtualisation of provisioning, organisation and interaction needed to happen, rather than the virtualisation of the storage itself. The ViPR controller does all of the array resource management. ViPR also offers a data service that will (eventually) provide file / block / object / etc translation services as required.
John then goes on to talk about Pivotal for a little while (the third bit of advice on how to keep pace with changing IT requirements). It can be boiled down to four things:
- Cloud abstraction layer;
- Big data;
- Fast data (real-time ingestion); and
John tells us that EMC believes in choice, which is why they’ve kept EMC, VMware and Pivotal separate (they can then sell us any of them or all of them). John then wraps things up and we move on to the EMC Galaxy Awards, presented by Shaun, John, Alister and Paul Harapin, VP APJ at VCE. There’s more about these on the ECN website.
Finally, the last part of the keynote is done by Michael McQueen – author, researcher and speaker – looking at what makes Generation Y tick. This was a fascinating and entertaining presentation, and I’d encourage you to check out his website and books.
And that was it for the keynote, a mere 2 hours later. I’ll be doing a post soon on some of the breakout sessions I attended.
*Update 16.08.2013 11:45am* I meant to link to Simon’s articles on The Register discussing John Roese’s position on capacity vs performance tiers – they are here and here.