As companies have been adding various types of storage devices to their networks, the task of managing an enterprise-wide system has been very labour-intensive, says Ken Steinhardt, director of technology analysis for Hopkinton, Mass.-based EMC Corp.
EMC manufactures storage-area networking
and network-attached storage products under brand names such as Symmetrix and CLARiiON, and under the AutoIS initiative.
Steinhardt, who joined EMC in 1994 after working for Digital Equipment Corp., says a major trend over the last 18 months has been the development of open storage management tools, which let users manage storage devices from different vendors and improve the way they back up data and allocate storage across the enterprise. He recently shared his perspective on the latest trends in networked storage and management with C&N.
C&N: There have been several different studies talking about the amount of money that organizations are spending on storage, and how it’s taking up an increasing portion of the IT budget. Do you think storage management is a very labour-intensive task, or a task which is getting more expensive?
KS: Historically, it has been extremely labour-intensive, and that’s what has been driving the need for storage management to be such a critical component of an effective IT infrastructure. Historically, people have had to deploy significant resources to manage not only individual storage products, but typically needed completely separate staff expertise and separate tools to manage the different storage products from different vendors. It was sort of, ‘Well, if I deploy multiple heterogeneous storage, that could get me certain advantages,’ but then they would lose them by having different tools, different management, different infrastructure and so on. Where we’ve seen a lot of impact in the last year and a half has been that customers have been able to move to open storage management. That gives them the ability to have common tools, to have common procedures, but then even more importantly, to have much greater functional capability than they ever had previously — to be able to allocate storage, make changes, back up information, to be able to have common disaster recovery or business continuity. We’re seeing significant positive business impact for customers — specifically, examples of customers who have been able to do much more creative things now, with intelligent software, than they were able to do previously on single vendors’ platforms now able to be even more productive with larger capacities across multi-vendor hardware platforms.
C&N: Can you comment in general on EMC’s strategy with respect to being able to connect systems from different vendors?
KS: We initiated a program a little over a year and a half ago, where the initiative was called AutoIS. The products that fall out from that give us the ability to do all of the fundamental management functions of storage and storage networks, with common tools across pretty much all of the major storage platforms in the market. The degree of functionality tends to vary, depending on the individual platform. The functionality tends to run the gamut, from being able to monitor health of individual multi-vendor devices, to being able, in some cases, to configure.
C&N: Some of the storage-area networks would be built around wide-area networks. What sort of network-related issues do you see , specifically with respect to the fact that it’s not going to be your own network? If you’ve got a couple of locations 10 kilometres away from each other, you’re going to be depending on someone else’s network to connect different storage devices, and maybe connect one storage device to your management console.
KS: One of the things we’ve tried to focus on at EMC is to stick to our core competency, which is storage. We’ve tended to work very closely with all of those companies that provide the hardware and software components that bridge off to a WAN for any functional capabilities that are specific to those things that would integrate directly to what we do with storage. For example, one of the more common applications would be remote replication of data. If you’re going not just ten kilometres, but thousands of kilometres, off to a remote site, we have tended to work with the individual third-party companies that provide those wide-area network technologies, to go through full test and qualification processes with our products.
C&N: I guess to a certain extent, the issues of uptime and reliability aren’t really storage-area networking issues. It’s more an issue of having the right service-level agreement with your carrier?
KS: Well, actually, they can be. In any chain, you’re never any stronger than your weakest link. I would say that you definitely do need to look at high-availability features pretty much at every single level, from the original client interface, into whatever local connection they make, through servers, through both local-area or wide-area networks, through storage-area networks, and ultimately to the storage and then back up again through the entire process. Without that, there can be vulnerabilities at any level of your infrastructure.
