IT Management

Solomon said a good name is more valuable than great riches.

Through the years, vendors have been challenged to come up with good names for computer products, and frequently change names of products that have decade-spanning lifecycles. IBM has certainly done its share of re-naming, even re-branding all four server product lines in 2000 and again in 2005/2006. When a system’s architecture receives a major extension, name changes intended to help make distinctions often leave customers—and even vendor employees—confused about what is what.

Sometimes names are assigned by engineers and sometimes by marketeers. The former tend to be systematic or scientific (i.e., logical), while the latter are hoping to be seismic (i.e., earth-shattering). To further complicate matters, the secret code names assigned to products during development may become so popular they never die, obscuring official names. An example of this was the “Shark” disk. IBM and customers had a lot of fun with this technology, whose official name was Enterprise Storage System (ESS). Plastic shark fins actually were seen on top of the devices in data centers.

The server marketplace is grouped into three general categories: small, midrange and large, and the high-end of one market can overlap with the low-end of another. Since the mainframe executive typically runs a multiplatform shop, beware of sifting through the nomenclature looking for a pattern or a progression. For example, since IBM has three products that compete in the lowend of the large-server market, some incorrectly assumed the AS/400 and RS/6000 must have been a step up from an S/390. It’s important to note in the largeserver, high-end space the mainframe still dominates because the breadth of its highly refined architecture allows customers to run mixed workloads and sustain high-transaction volumes in a reliable (near zero downtime), secure (free from malware), and scalable (deliver consistent response times during peaks) fashion.

The current mainframe product line spans more than 40 years with an evolving system architecture that debuted in the ’60s as System/360. S/360 boxes were general-purpose computers and the name comes from the 360 degrees in a circle. The System/370 followed in the early ’70s, featuring storage virtualization. In the ’80s we saw System 370/XA (eXtended Architecture) raise storage addressing limits, and System 370/ESA (Enterprise System Architecture) with sophisticated processor virtualization. The evolution continued with System 390 (’90s), zSeries (2000), System z9 (2005) and System z10 (2008). In the re-branding of System 390 to eServer zSeries and then System z, the z stands for zero downtime. In the 2000 re-branding, IBM’s other three eServer families were designated the pSeries (POWER Architecturebased Unix machines formerly RS/6000), iSeries (integrated servers formerly AS/400s), and xSeries (extending hardware function for systems built with Intel processors). Today, these are branded System x, System i, System p, and System z.

In recent years, IBM has become much more intentional about sharing hardware technology and software across product lines. For example, features refined for decades on the mainframe (e.g., virtualization and workload management), to varying degrees, have been cross-pollinated into the other server families. IBM has invested billions to make Linux a common operating system that spans all four server lines. I can testify to the legitimacy of IBM’s Linux strategy. It took me and two colleagues only 15 minutes to get our software products (online exam delivery applications) installed and operational on all four IBM server families (z, p, i, and x) because our software is Webbased and runs on Linux.

A few years ago, the pSeries and the iSeries lines converged to the extent of sharing the same POWER architecture hardware and firmware base. This architecture is in its sixth generation, or P6 (thank you, engineers!). Some aspects of P6 technology have been leveraged in building the z10, which is why the z10’s hot new quadcore chip was dubbed the z6. Beyond that, convergence simply isn’t practical because System z architecture has a number of things, such as its complex instruction set, that are far too sophisticated to shift to POWER. The important thing for an executive to remember is that mainframe technology, by any name, is still the gold standard by which all other platforms can measure themselves. It doesn’t really matter if the name sizzles, as long as the platform does.