Last week, IBM continued to move on its On Demand computing strategy by releasing its much-awaited Storage Area Network (SAN) TotalStorage product. In addition, IBM also announced that it would purchase CrossAccess Technology, a privately held company that specializes in data integration software for mainframes.

How do these announcements fit into IBM's grand scheme for an On Demand computing infrastructure? What does this infrastructure mean to the iSeries community? While still years away from realization, the On Demand environment will arrive by small technological advancements that will change the way our organizations function.

Storage Area Networks: IBM's Is First

On October 13, 2003, IBM announced the availability of TotalStorage SAN. TotalStorage SAN is the industry's first storage software technology that allows companies to easily share millions of data files in a heterogeneous environment while simultaneously enabling the automated management of the data. Today, this is a clearly defined mainframe solution, but these services will be extended down to the iSeries, pSeries, and xSeries platforms to pull them into the SAN architecture.

The SAN ties together servers in multiple locations across an Intellectual Property network, allowing the network to operate like a single local file system, regardless of the location of the data or the operating system of the serving computer. It provides a single, centralized point of control to manage the storage devices and the data.

This simplifies the administration of diverse systems on disparate platforms by creating a new kind of "virtualized" software layer. It also tracks the descriptive metadata of physical location and file size and maintains the original levels of access permissions.

The SAN file system is built around IBM's grid and autonomic technologies using an architecture that can ultimately support thousands of computers, petabytes of data, and billions of files. (A petabyte is 1,024 times larger than a terabyte.)

This TotalStorage SAN file system technology is currently being used at the John Hopkins University, where researchers are developing a global file system to service dispersed research organizations around the world.

Relating to the Technology of SAN

Today, in our iSeries-centric vision of the universe, we experience the challenges of sharing information as either a "data interface" problem (for instance, installing a JDBC API to gain access to a database) or a "client-server interface" problem (for instance, building the connection profile to access a remote server). There are thousands of different kinds of interfaces, each requiring their own unique tweak or configuration. But this is really a kind of computer "plumbing" that results in increased levels of complexity and requirement maintenance that inevitably has problems.

The concept of a storage area network, such as the TotalStorage SAN, attacks the problem from a completely different angle. It offers a new level of virtualization to all file and storage systems by building a software layer that resides above the operating system and above application program layers of the architecture.

On the iSeries, this is similar to the IFS, pulling together various file systems from PC, UNIX, Linux, and OS/400 operating systems into a single structure. The SAN provides the protocols and the permission layers that describe a complete network of files that reside virtually on a computing grid, making the file resources of all the connected computers appear as a single layer of service. Tied together by broadband, the SAN enables a common industry-standard interface into which data resources can be plugged, and by which monitoring, access, and maintenance can be applied. It removes the complexity of disparate interfaces to make the files appear in a single level of storage.

The TotalStorage SAN removes the first major technological obstacle that has prevented large organizations and governments from collaborating more effectively: the physical location and connection of disparate computing systems. And while the current TotalStorage SAN offering is still well beyond the means of most corporations, it's fast approaching iSeries shops that have parent companies running mainframe networks.

From File Share to Data Integration

In related news, IBM's announced acquisition of CrossAccess' DB2 integration technology. This technology will allow IBM to deliver a more comprehensive enterprise information integration infrastructure. According to IBM, no other software vendor provides the depth and breadth of this kind of database integration.

Why is this so important to mainframes?

In the mainframe environment, there are numerous legacy data environments, including IMS, VSAM, CA-IDMS, and others that create major headaches for integration of information. The technology offered by CrossAcess transforms these disparate data environments into an intermediate e-data environment, so that the information can be accessed easily through Java applets. IBM will use this technology with its Data Integration offerings to enable more comprehensive supersets of data to be assembled.

From the iSeries-centric perspective, such data integration requirements may seem obscure, but for organizations that routinely try to interface their iSeries users to mainframe environments, the implications are significant. With so many different kinds of data environments--each requiring different applications to access the core information--data integration to some common platform has, in the past, been a major challenge. APIs and services (such as Web services) are often seen as the only practical tools by which to pull these disparate hunks of information together. However, each of these kinds of solutions brings its own level of complexity and challenges.

IBM has been building a mainframe DB2 Data Integration product to simplify the integration model. Now, instead of attacking the problem from the application layer--providing a unique interface to each application--the CrossAccess technology will allow IBM to build an intermediate eData layer that will allow a common access level of data integration.

Impact to the On Demand Model of Computing

Why are these announcements important to IBM's On Demand computing model?

One of the most important goals of IBM's On Demand model is the transformation of companies and organizations into more effective work units through the increased consolidation and collaboration of information services. The On Demand model envisions the breakdown of silos of authority within organizations and across industries by providing common processes and common services through the entire breadth of the traditional hierarchies. This is not a hardware vision of the future, but a real transformation of organizations and the IT services that support them.

For instance, in a supply chain, the individual organizational processes that separate a supplier from a manufacturer will be wrung out in order for increased efficiencies in producing products for a global economy. Just-in-time inventories will be transformed in this economy to "just-enough" inventories, to reflect real-world demands for products.

To allow this to happen, closer collaboration must occur between the supplier and the manufacturer, and the IT services that support these organizations must reflect those requirements. Right now, IBM estimates that 60% of all business done in the global economy is performed by mainframe computers, and the ratio is expected to increase dramatically.

A Mainframe Universe?

Does this mean that we'll all end up on mainframes? Not at all! But it does mean that initiatives like UCCnet, which creates central registries of universal part numbers, will become the status quo. And this, in turn, means that IT will be required to establish a kind of universal set of IT services that consolidate IT infrastructure and reduce IT complexity.

We are already witnessing the transformation of the manufacturing industry along global supply chains, and this is creating tremendous pressure for companies to consolidate and integrate their IT services to meet the demands of efficiency.

IBM believes that, in the long run, the IT services will be seamlessly integrated into a vertical chain of demand--from customer through manufacturer to supplier. But to facilitate this kind of efficiency, there must be a horizontal integration of collaborative services through common processes and common services.

Vertical Integration and Horizontal Collaboration

For instance, in IBM's vision, customers and suppliers will be brought into the engineering design phases of new product development, so that the products actually created will be more efficiently engineered and will more accurately reflect what is required. In order for this kind of collaboration to occur, the traditional silos that prevent this kind of collaboration must be removed.

This includes disparate IT architectures supporting separate purchasing departments, human resource departments, marketing departments, production departments, etc. According to IBM, these separate hardware systems (if they are currently separated) must be brought into a vertical integration along the lines of international standards, allowing for consolidation for better efficiencies.

Once consolidated, individual industries will create common communication processes along burgeoning international standards and registries. (UCCnet is the current focus in the manufacturing and distribution industries.) Relieved of the technical complexities that disparate IT services require, IBM believes organizations will ultimately arrive at a point where most of the infrastructural IT services themselves could be purchased on demand, much like the utility services of gas and electricity.

Where Does the iSeries Fit?

In this light, the current configuration of the iSeries represents a small-scale image of how the entire on-demand infrastructure might appear in a global environment: The iSeries connects to just about everything, allows server consolidation of dissimilar operating systems onto a single box, provides an excellent infrastructure to reach outward to the global supply chain, and has some excellent on-demand capacity management facilities that are, in many respects, heads above the competition. It is a "virtual architecture" in a nutshell of electronic circuits and scalable code.

However, in IBM's larger scheme of a global On Demand computing environment, the iSeries is just a small oasis of sanity in a significantly larger terrain of complexity and conflicting standards. No one imagines the world populated by iSeries anymore; instead, they imagine a world in which the services that are the hallmarks of the iSeries extend to a virtual computing environment around the globe.

It is this image of on-demand computing--in which IT services are invisibly and virtually connecting business relationships--that allows the iSeries to represent the future of computing that will transform how businesses and organizations will operate in a global economy. And in this light, IBM's TotalStorage SAN and acquisition of CrossAccess data integration technology make perfect sense to the larger scheme and challenges that IBM faces. While we on the iSeries platform have been spoiled by the platform's level of integration and transparency, the rest of the world is starting the journey to catch up with us.

Thomas M. Stockwell is Editor in Chief of MC Press, LP.