The coming IT integration hellhole

In a futile attempt to fend off the PC invasion in the Unix workstation market in 1989, Scott McNealy, then CEO of Sun Microsystems, convinced Lotus Development Corp. and Ashton-Tate, two of the largest independent software vendors at the time, to port their famous spreadsheet and database applications, respectively, to Sun’s fabled workstations. Little of note came from the deal beyond a few misspent millions of dollars in development. Additionally, by the end of that same year Sun was sitting on more than 10,000 unsold Sun 386i workstations, the company’s failed and very expensive attempt to merge the engineering workstation with the desktop PC.

Despite the obvious similarities of networked, desktop computers, workstations and PCs did not serve the same business needs, which was why Sun’s attempts to bridge the two failed. It’s also why through the beginning of this century, you often saw certain workers with both a Windows PC and a Unix workstation on their desks.

There are some unnerving similarities between software as a service versus the packaged and custom application software markets of 2009 and the PC business of 1989. One obvious parallel is that the software people got on PCs in 1989 was available on mainframes, minicomputers and even workstations at the time, just as SaaS tools are available through other means today. But in 1989 DOS-based PCs became extremely popular with individuals because they could run spreadsheets, create simple databases and crank out memos without having to beg CIOs (or MIS managers, as they were known back then) for resources, so they snuck PCs in the back door, as it were, department by department.

In 2009, SaaS is in a similar situation. End users can quickly get access to sophisticated applications at a low entry price-point. They do not have to procure, install and manage packaged products. By and large, users can bypass the CIO to get what they want. As a result, lots of companies are seeing SaaS slip through their doors via the marketing, sales, HR and finance departments. Eerily enough, Microsoft has even been reprising Sun’s attempt to blend PCs and workstations with its software + services approach to SaaS.

The ad hoc approach to embracing the PC led to numerous problems that all ended up on IT’s lap. Clashing standards in everything from LAN protocols (token ring vs. Ethernet) and cabling (twisted pair vs. coax) to incompatible applications (1-2-3 v. Excel; Word Perfect vs. Word) wasted everyone’s time and money. Companies like Bridge Communications Inc. emerged simply to handle the translation of packets from one network to another; not unlike what Boomi Inc. is doing for SaaS integration today.

It took years, decades even, for PC users and vendors to sort things out so you didn’t need a middleman (or middleware) to make things work together. So, will today’s piecemeal approach to SaaS create similar issues for businesses down the road? It already is. Will CIOs be devoting huge portions of their budgets to bridging the gaps between departmental SaaS projects and corporate-wide application standards? Undoubtedly.

SaaS vendors will argue they use existing standards, such as Web services, that will obviate the problem. In 1989 PC hardware and software suppliers said the same thing, though back then it was called IEEE 802 and CORBA. But the emerging integration hellhole was still immense, complex and expensive, and in many organizations lingers to this day.

 As Loraine Lawson points out in her blog, it would be foolish to assume that history won’t repeat itself. So prepare yourself for another long period of eager user adoption and inevitable IT headaches.