How the Cloud Can Lock You In

Saturday, February 05, 2011

Rahul Neel Mani

F520f65cba281c31e29c857faa651872

Ken Steinhardt, VP, and CTO, EMC, talks about the closed nature of public clouds and how eventually data migration would get non-disruptive. Steinhardt spoke in an interview with Varun Agarwal.

Q: What do you think 2011 would be about? What kind of evolution do you see in virtualisation technologies going forward?

A: To me cloud is the next logical evolution of visualization. The first wave of visualization to me is when people virtualized in those four traditional tiers of IT infrastructure – Storage, network, Server, and operating system.

To me cloud is nothing more than virtualized infrastructure and at the same time, I don’t want to know what the server is, what the network is, what the storage is. I just want to know if there is a place I can run my applications and store my information and have pre defined service level.

However, visualization has been there since the mainframe era and the flexibility and the barriers to change today is so different and so much lesser than they were back then. The barriers to change back then were that the applications and databases tended to be unique to individual server platforms, they could never run on anyone else’s load.

But today through standardization, software and the interconnects, the barriers to change are nowhere near what they were back then, so it’s a lot easier to get from one place to another.

I will give you a bold prediction — in over the next couple of years, some of the storage visualization technologies are going to see the ability for non disruptive migration of data from any major vendor to any other major vendor without having to shut down anything.

So you will see the ability to federate multiple heterogeneous devices- servers, storage, networks — where people will be able to replace the infrastructure components individually without shutting down the others.

This will be a fairly revolutionary breakthrough. Historically people have had major issue with migration sometimes driven by the fact that they have to disrupt production environment to be able to make changes.

I personally believe the greatest single either enabler or obstacle towards cloud infrastructure is security. What I think a lot of people miss sometimes is the security from all perspective. Organizations look at it as I have to protect my data either in terms of importance or for compliance reasons, maintain it and have it totally secure.

Individuals feel it the same way, no less important to them whether they are storing family photos or personal information, the security needs to play at both ends so whether it be the aspect of authorization, authentication, or encryption of the data, information management security becomes a key component.

Q: Do you also see public clouds being neutral completely wherein you can shift your data from one cloud to another?

A: Great point. This has become a huge obstacle today. A particular public cloud infrastructure can be very much of a lock in today, which can be a problem. What the industry needs there is some standardization, which from a cloud perspective there are some discussion going on in some various bodies now but there is no technology today to standardize making that seamless for customers yet.

Once you're in the cloud, it can get difficult to get out of it. I believe there will be people that will need to go in both directions and transition from traditional infrastructure to cloud has to be easy. There will be some that will realize may be the cloud infrastructure wasn’t at that point in time the appropriate solution for them, they didn’t have the flexibility to make it easy to move out of cloud. These are the kinds of things where non disruptive movement of data between heterogeneous devices becomes a key.

Q: EMC has been on acquisition spree for quite a while. What would be the technologies or companies you would be looking at?

A: We have pretty consistently stuck to the ones that are close to where our core competencies are. I don’t believe we will stray from that in future acquisitions. Virtually all the major acquisitions we have made for fairly extended period fell into one of the four categories: security, document management, information management and visualization.

Then we tended to look for the company to be the prime integrator within that category so for security it was RSA, for document management it was Documentum, for visualization a lot of it was driven by VMware. Now we realized to advance the category and we went to the consumer space with Iomega and with Mozy in cloud space. That evolved to be our new plan.

These are really new interesting categories where we must play. All our acquisitions methodically pretty much revolve around these, I can’t think of any significant exception into those primary categories plus the two which is basically cloud and consumer.

Q: Any specific technology EMC has or would really be bullish about?

A: There are a couple of technologies right now that we're really bullish about. first is the concept called fully automated storage tearing (FAST) which is the ability for the storage system instead of having to be a specific tier i.e. high end or mid tier, the ability to typically have different class or tiers of storage technology (each with a different associated with it) all co exist on one system at the same time. The vast majority of applications have a very wide skew between the number of I/OS that run on a particular volume and the number of volumes.

This is what we found, is usually it's the tiny number of volume generates a majority of I/OS, there is usually a large number of volumes that generate with very little I/OS then there is something in the middle. So logically if you can move just a little bit of the hot ones that were hot up to flash drives to almost everything else to lowest cost drives, we bring massive reduction to space, power, cooling cost complexities but you actually improve the performance.

Historically those two things have been diametrically opposed — I want higher performance that means I am probably spending more money for more expensive type of device I want to reduce my cost I compromise my performance and service level.

This is being able to do both at the same time so the customers that have been deploying these technology had it in the market for about a year now could basically have consistently been saying massive reduction in cost space power cooling by configuration are actually shrinking or my performance is improving while I do this.

The other is de-duplication. I believe that virtually all environments at source, at target would benefit over the next time especially over the next few years with de-duplication. up until now if we were to save the same PowerPoint file, systems were designed to do what they were asked to do, i.e. every single instance of it, every single copy I store, is taking up unnecessary additional space. So being able to go into de-duplication becomes key. Now we have been massively acquiring de-duplication intellectual capital for EMC like Avamar and Data Domain most recently.

The couple of other of basic ones that are becoming pretty common in the industry, but I will mention them and just because they are becoming common doesn’t mean they are any less important – thin provisioning, the premise of thin provisioning has a very basic value proposition: one its unbelievably easier and faster to provision and other, is direct impact on capex.

Cross-posted from CTO Forum

Possibly Related Articles:
10659
Cloud Security
Storage Cloud Computing EMC Information Technology Enterprise Data Migration
Post Rating I Like this!
The views expressed in this post are the opinions of the Infosec Island member that posted this content. Infosec Island is not responsible for the content or messaging of this post.

Unauthorized reproduction of this article (in part or in whole) is prohibited without the express written permission of Infosec Island and the Infosec Island member that posted this content--this includes using our RSS feed for any purpose other than personal use.