Among those who manage storage in large, heterogeneous environments, it is no secret that understanding the relationship between various IBM storage management products can be a confusing endeavor. This may finally change in the near future, however, as all storage management products from the TotalStorage and Tivoli Storage product lines will likely emerge under a single product umbrella name.
But it is not just about changing names. Rather, it is an effort to ratchet up IBM's On Demand and autonomic messages relating to the subject of enterprise storage in open, heterogeneous environments. "[IBM TotalStorage and Tivoli Storage] is really one On Demand story, so therefore having the two brands...creates confusion," says Jens Tiedemann, Vice President of Storage Software, IBM.
Already the development teams of IBM TotalStorage and IBM Tivoli Storage have been working more closely together. Each group has its own development organization, but they will be governed by one investment review board. Says Tiedemann, "We are now looking at what to invest in at the same time in the same meeting with people from Tivoli and TotalStorage."
It also wouldn't be surprising if the sales teams were combined at some point to better present a cohesive storage management message to customers. Continues Tiedemann, "[Customers] could have one guy in the door one day talking about On Demand (relating to TotalStorage products), and the next day some person from Tivoli would have the same story, but it wouldn't be connected to what they just heard. We have not spoken the same language....It has been difficult for the customer to make the distinction, to figure out what to buy where and when."
Another issue with IBM's communication about its storage products is the lack of emphasis on "openness," or the cross-platform compatibility of its various storage management products. Considering that enterprise storage environments are not usually made up of one brand of storage hardware, this is certainly something that needs to be addressed.
Expect announcements soon from IBM that address many of these issues and work to focus all storage products under one umbrella, all communicating the same message.
Virtualization Technology Is the Key
Perhaps the foundational technology of the IBM On Demand and autonomic messages for storage is virtualization. In fact, many other IBM storage management technologies (e.g., hierarchical storage management, policy-based rules, provisioning, etc.) work far better when virtualization functions are in place. Therefore, it just makes sense that all storage products eventually plug neatly into one family of products if IBM is to clearly communicate its value propositions and build a development model that can be more easily understood by customers.
For those who may be wondering exactly what virtualization is, let's digress for a moment. Virtualization is essentially defining a collection of diverse storage resources--typically connected together in a storage area network (SAN)--in a more simple, logical, and manageable way. For instance, instead of managing hundreds of disparate storage resources separately, some or all of these resources can be pooled together into a single "virtual" storage resource that is managed by a single storage controller. Think of it as taking a hundred different connected storage devices and putting them all in one box with a single controller. With virtualization in place, a typical large computing enterprise can save enormous amounts of time and money as the job of managing these resources is infinitely simplified.
Without virtualization, it is the server that has to keep track of where its data resides on all connected storage resources. With virtualization in place, the server thinks its data resides on a single storage resource ("virtual"), but the virtualization hardware and software do all of the work of keeping track of where the data resides on the plethora of connected storage resources. On its own, the virtualization storage controller moves the data around to the storage devices that it deems best, depending on hierarchies and other policies that are tied to events usually defined and managed by various IBM storage management products. New disk space can also be automatically "provisioned" when and if it is needed--again, based on predefined policies. Because this underlying virtualization layer manages the data storage automatically, the operator no longer needs to manually make many of the daily storage management decisions. Even more importantly, however, downtime that normally results from many different system maintenance events can usually be eliminated, often saving the company many thousands of dollars per event.
Down with Downtime
Virtualization not only simplifies the task of redefining, adding, and replacing storage resources (which on its own saves enormous amounts of money in reduced operator time), it also dramatically reduces or eliminates downtime that is usually associated with these tasks, as well as unplanned events such as hardware failures.
Without virtualization, attached storage is literally connected to servers through a physical "umbilical cord." This means that when storage needs to be added, reconfigured, or removed, the server must be taken offline. With virtualization, the physical link essentially becomes irrelevant, as all controlling of storage is done within the virtualization layer. All that the server knows is that one storage resource is connected and it is always connected, regardless of what is happening to the individual storage resources that are connected to the virtualization layer.
Are you starting to get the picture of how virtualization is a pivotal part of IBM's On Demand and autonomic message for storage? With this, it only makes sense that the computer giant is finally realizing that it needs to communicate about virtualization (and the myriad of other storage management technologies that benefit from virtualization) with one voice.
The Problem of Over-Purchase, Under-Allocate
Efficient storage allocation is another huge benefit that comes with virtualization. Without virtualization in place in SAN environments, managers commonly over-allocate disk space because they want to make sure they don't have to interrupt users again for a long time to reallocate more disk space. This of course translates to low average disk utilization rates. In fact, in a Q4 2002 IBM SAN Total Cost of Ownership study, it was discovered that the average SAN has less than 50% of the disk space being utilized, which translated to an average of 1 TB excess capacity per SAN! With virtualization in place, disk space is not only allocated as much as is needed without creating downtime, but this process can be done automatically based on provisioning policies that are predefined in various storage software products.
Currently, the features that either create virtualization or create the myriad storage management benefits from virtualization are available within two different storage management product lines within IBM. One is the IBM TotalStorage Virtualization Family of products announced last April, which includes the IBM TotalStorage SAN Volume Controller (announced last summer) and the IBM TotalStorage SAN File System (aka "Storage Tank"). The other product line comes from IBM's Tivoli side, which primarily includes IBM Tivoli Storage Resource Manager, IBM Tivoli SAN Manager, and the widely used IBM Tivoli Storage Manager.