EMC is promoting Automated Information Storage Management (AutoIS) – its strategy for reducing management complexity in a multivendor, networked storage environment – with great fanfare. AutoIS shows that EMC has a grandiose vision for the future of storage management. EMC’s WideSky technology is a key enabler of AutoIS; however, we believe that enterprises need to be concerned about using WideSky as a storage management middleware.
Background on AutoIS
AutoIS is EMC’s vision for storage management. As EMC moves more of its efforts into software, it is attempting to use its large installed base to leverage more software sales. AutoIS is a vision that includes heterogeneous storage device management via a management framework with a set of application programming interfaces (APIs), a database and a graphical user. Enterprises would greatly prefer storage management solutions that are not dependent on the underlying hardware and ensure heterogeneous device support regardless of who delivers the software (for example, there is no penalty for a hardware vendor to provide software that also manages the competition’s hardware). This is why, conceptually, AutoIS makes a lot of sense and has received lots of positive user feedback.
However, enterprises need to recognize that it is relatively easier to create a vision in storage management than it is to execute it, which involves improving the quality of service via availability, performance and provisioning, while continuously lowering costs in a completely heterogeneous environment. The key caveat here is that there is a wide gap between vision and real software implementations. In fairness, EMC has announced the vision and is building products to close this gap; however, it is unlikely that these gaps will ever be completely filled in. EMC believes that delivering on this vision will require a management framework that hides the nuances and differences between devices. This is where WideSky comes in.
What Is WideSky?
WideSky is a proprietary management framework. Frameworks have a long history in network and systems management (NSM). They have experienced a lack of success to date in mainframes and distributed computing because independent software vendors (ISVs) do not want to be dependent and controlled by other vendors’ software. In addition, framework vendors have spent too much time developing and maintaining their framework while neglecting the NSM applications on top of the framework.
Furthermore, hiding the semantic differences of the managed objects limits ISVs to treating all of the managed objects the same, when the objects themselves are being purchased because of these differences. All frameworks have been proprietary in nature and have been specifically designed to create further account control with customers. With WideSky, EMC is trying to control the future of storage management, with only EMC deciding what is included in WideSky; what APIs, functions, language bindings, software services and databases are supported; and when any of these possibilities get implemented.
Therefore, enterprises should not expect WideSky to become a de facto standard for the storage industry.
Enterprises should expect EMC to claim victory. Because there is a large installed base of EMC hardware, ISVs that need access to these hardware APIs are forced to use WideSky (or usage of these APIs will be construed as using WideSky). At last count, 184 API licenses had been issued. This claim of success should not be viewed as the creation of a proprietary industry standard, but WideSky did provide a catalyst for EMC’s competitors to promote the Common Information Model (CIM)/Bluefin standard (even before there was any substance to Bluefin).
There are three possible outcomes to the evolving efforts for storage management standards:
- By 2005, WideSky will become a de facto standard for storage management that is controlled by EMC (0.1 probability). The reason this could occur is because ISVs are forced to use it for access to EMC hardware. This is very unlikely because ISVs hate to be locked in and cannot be reliant of EMC to address future enhancements, support or additions. ISVs will also find a way not to pay a royalty fee for every product shipped.By 2005, there will not be a usable industry standard for storage management (0.2 probability). This could be the case because of politics among vendors and no vendor having a sufficiently dominant market share.By 2005, a storage management standard will be developed through an open, industry standard consortium or organization (0.7 probability). With all of the work occurring with CIM and XML, as well as the mass of vendors supporting the Storage Management Initiative from the Storage Networking Industry Association, this is the most likely outcome. However, standards in all of the management areas have been extraordinarily slow to date.
A Standard for Storage Management Is on the Horizon
With the advent of CIM as a standard for management in many areas, including storage, there is a standard for storage management on the horizon. However, standards have historically been slow (this is certainly true of CIM), and standards only provide the lowest common denominator for management. Any standard – CIM or otherwise – will be insufficient, and there will be proprietary mechanisms to deal with new storage innovations as well as legacy devices. This will create proprietary and semiproprietary extensions (proprietary, but in CIM format). EMC has committed to supporting CIM in 2003, along with all other storage vendors, and is building CIM support into WideSky as well as extensions beyond CIM.
Building WideSky as a proprietary framework to lower development costs, testing costs and as a means for EMC to address heterogeneous storage management is a solid strategy for EMC software product development. With WideSky, EMC addresses its own internal software compatibility issues between the CLARiiON and Symmetrix product lines. EMC can build good storage management software and be successful as a heterogeneous storage management vendor even if WideSky does not become a de facto standard. However, EMC’s attempt to create an industry standard that is entirely under the control of EMC is not in the best interests of IS organizations.
For EMC to create an “open” industry standard, it must not charge for the runtime license or maintenance fee, and relinquish control of the software to a standards body or independent, nonprofit software organization.