I had the pleasure of spending my lunchtime at another software company this week – and was asked to be a judge in a “gingerbread house building” game that pitted developers against QA, support, sales, and management. The gingerbread houses were pretty impressive, and the photo I have included here is of me digging into a gift bag I was given after the results were announced.
And it wasn’t long before we began reminiscing about the old days, and about the applications we wrote many decades ago. Remember PL/1, I was asked? Remember teletype protocols and paper tape? Remember memory drums? While I have spent the last two years working remotely, and have enjoyed the flexibility it provides, I sure do miss times spent around the water cooler!
I wrote earlier this week an article for the upcoming December issue of the electronic newsletter, TandemWorld. In it I referred to an advertisement by IBM in the November 28, ’08 issue of the Fortune magazine. IBM, under a large-type headline “Decrease costs. Increase growth” opened with “Legacy apps. Faulty data. Redundant silos. In banks all around the world, millions of dollars are tied up in outdated, patchworked systems.” Without revisiting the topic and covering the same material, it is the reference to “legacy apps” in the above context that I find amusing, particularly coming from IBM.
Last year, in a posting I wrote while in Singapore “What do you mean, legacy?” (November 27, ’07) I recalled how “I once heard an analyst with Gartner tell their audience to remember that, any application that is put into production, should be considered legacy immediately!” The Gartner analyst may have been over-dramatizing, but in a recent email exchange with Jim Johnson, Chairman of the Standish Group, he followed with a similar line, remarking how “there was the old joke that a legacy system is one that is working!”
And then Jim took it a little further, suggesting “the term’s roots were used in a more derogatory fashion and it meant that the legacy applications were running on outdated hardware and software, most meaning IBM mainframes. People promoting the use of open systems mostly used it.” Legacy applications, it would seem, have as much to do with which platform was chosen to deploy them, and the infrastructure supporting them, as anything else and typically, this means dependence on much older proprietary Operating Systems (OS) and utilities. Jim then added “a truer definition might be a legacy system is a system handed down from a predecessor.”
In other words, the reference to legacy when talking about platforms started with the arrival of Unix and open systems as they gained momentum in the late ‘80s. I suspect that it was Sun, and most likely someone like Bill Joy, who first threw down the gauntlet to the more traditional vendors admonishing their user communities over maintaining legacy applications and holding onto their legacy equipment and infrastructure.
HP was certainly among the front-runners in adopting Unix and the HP-UX that appeared in the early ‘80s had it’s roots deep in the early releases of Unix – System III and later System V – and some three years before IBM released AIX/6000. I recall attending early presentations given by HP in ’84 where it was clear that they were making Unix their main operating system, and that it heralded a break from other vendors going down the IBM Plug Compatible Manufacturer (PCM) path that was in full swing at the time. The legacy net, cast as it was over the IBM and IBM PCM community, caught a lot of vendors!
But the legacy label is still a hard label to shake off. For companies that have been successful for many years, and with product roots going back to the ‘70s or even earlier, it is particularly hard to distance the modern architecture they have today with what preceded it. Nowhere is this more recognizable than with the enterprise class of systems now being shipped by IBM and HP.
While it’s hard to consider an IBM System z running mainframe Linux or HP’s NonStop running on industry-standard blades packages and supporting an open-source infrastructure stack as legacy, many CIOs continue to overlook the contributions these platforms can make. Because these vendors have striven hard to make sure that everything developed in the past can continue running (and further entrenching the thought of legacy) – should this then penalize their current product lines in any way? And the emergence of a service-oriented architecture in support of users scattered across the globe really does beg for servers as available as NonStop.
As I talk to the vendors, it has been my observation that as we move from distributed computing to grid computing, and onto cloud computing, having systems such as these protecting our data bases, for instance, wouldn’t be all that bad an option. Moreover, when the time comes for rationalization or consolidation following inorganic business growth, after an acquisition for instance, which is happening a lot these days, anchoring key business data on a system of this type becomes a serious consideration.
The December 1st, ’08 issue of InformationWeek features an in-depth article on HPs CIO Randy Mott, and the transformation of HP’s IT that he has overseen. While the author was impressed with the accomplishments of Mott, he asked him how concerned he had been about this far-flung network of legacy applications? According to the author “Mott points, as an example, to HP’s networking and telecom costs, which were driven by having to connect those 85 data centers around the world and hundreds of data marts to share information. ‘A miracle had to occur every day’, (Mott) says of the complexity. ‘The bad news was that a miracle never occurred.’”
I wrote about Mott’s transformation project a year ago, and in the posting of November 11th, ’07 “We all just wanna be big rock stars!” I included a quote from the Financial Times of November 7th, ’07. Under the headline “Techie rock star’ sets a cracking pace”, the author reported “Randy Mott’s trail-blazing project to overhaul HP’s technology could set the standard that other companies need to follow …” I then added “just as successful rock stars have changed the course of music and carved out highly lucrative careers, so do some CIO’s as they press ahead with innovative ideas.
Having the data base at the very heart of Mott’s new architecture running on NonStop certainly doesn’t appear to be a backward step for HP and certainly doesn’t appear to suggest that the platform was considered legacy. According to the author of the InformationWeek article, “Mott insists he wasn't forced to use Neoview, but let's assume the corporate pressure was at least implicit. He says the technology team was told to evaluate whether HP had a marketable, scalable product, and it worked with the development team to fix any bugs. The result is a product built around the Tandem NonStop OS and database. Mott says HP's will be one of the largest data warehouses in terms of size and number of users, and one that will ‘not go down.’”
In a recent email exchange with a senior NonStop development manager, who agreed with the sentiment behind the earlier definitions of legacy, explained to me how “we have worked very hard over the years to provide forward compatibility, including completely re-implementing the preexisting Guardian callable procedures so they would continue to function correctly within existing limits when we did the Exceed version of the operating system 15+ years ago. (There’s) a vague recollection that we made an incompatible change (for at least some programs) to the object file format somewhere back in the early 1980s, but other than that I don’t know of any reason why an ancient object file wouldn’t run on an Integrity BladeSystem.”
It was a lot of fun joining a software development company this week – and in talking revisiting our exploits from the past. There’s still no substitute for experience and just being able to talk about the history of technology with passion was enjoyable. And it gave me time to appreciate the distance we have come with NonStop and to realize that if it continues to leverage modern hardware packaging and support current development environments, it’s as valuable today as it’s ever been.
If you ever thought about what legacy your grandparents left behind … you probably did not think about it in derogatory terms, eh? Perhaps Jim Johnson in is final email exchange with me has it right after all “in today’s complex and service oriented environment the word “Legacy” is more of a legacy than a current term with any real meaning!”