Legacy Systems: The Heir Bear Bunch
Guy Clapperton reports on the support and hardware that is available to companies updating their legacy systems
A long time ago, and in a universe far, far away, a family friend wanted to upgrade her old computer. Being young and impressionable, and at the time really naive about how helpful the computer industry would be in revamping an existing system, I gave her the manufacturer?s PR contact.
Within days I received what should have been a smiley telephone call. ?Thanks,? said the caller through gritted teeth. ?So do you know anyone with an abacus we can upgrade??
There are definitely times when a customer with an older system is, to put it frankly, a bit of a pain. At the retail end of the market, dealers will tell you with a distinct shudder about the person who comes in wanting software for their Apple 3 system, or their Amstrad word processor or Z88 laptop. Quite rightly, dealers are politely unhelpful and try to persuade these people to buy something modern and upgradable.
The snag is when Johnnie punter isn?t a lone bloke with a creaky 8086 on which he wants to install Windows NT, but a full-blown corporate with a Vax, AS/400 or mainframe on which it would quite like to run an intranet, if that?s OK. There are ways around this, and profiting from the requirement is manageable. Whether it is easy or even desirable depends a lot on the sort of organisation and skillset you are offering.
For a start, the manufacturers themselves are pretty determined to make things easy for people who want to upgrade rather than replace. Take IBM?s recent addition to the AS/400 family, the AS/400e ? the idea being to take advantage of the internet, Domino, Active X, all that modern stuff. And it will be fully compatible with an existing AS/400 system, so you can extend as well as start from scratch.
But selling one of these isn?t quite the same as selling into a complete older system ? technical skills are required as well as knowledge of a whole load of products and workarounds. A UK-only search on the Web yields literally hundreds of sites, each believing it offers a different and more valid set of products than the others.
If it?s a business area that interests you, a good place to start is to talk to a company that already specialises in it, such as distributor Unipalm. Product manager David Ellis believes there are a lot of resellers interested in the field. ?People have spent a lot of money on their systems and they want to protect that investment,? he says.
Setting up a reseller network has been no problem for Unipalm, which has been established for a fair while. It has been slightly trickier for the recently established Open Connect in the UK. The company looks at SNA-to-Web access and has had its product, Web Connect Pro 3, certified as 100 per cent Java compliant. But so far, Open Connect has only signed up three resellers, explains general manager Graham Jones. ?Currently we are more into the OEM channel through Cisco, Oracle and Sun,? he says.
The reasons for this are many, and mostly to do with the fact that the company specialises in connecting to mainframe architectures. The sales leads come from asking prospective customers whether they have a mainframe, and if so Open Connect asks to get involved. ?Modesty aside, we know the market,? says Jones. And the resellers, bless ?em, don?t seem to just yet.
Jones has begun setting up a network of non-OEM resellers, but he is finding it hard to match the quality of, say, the Cisco salesforce. Given that the Cisco staff will have access to IBM decision-makers at MIS level, this shouldn?t really come as any surprise. But the situation is not quite as clear cut as that. ?There?s no real expertise [in SNA] in the UK at dealer level,? says Jones.
Unless of course, you start looking at the Computacenter size of operation. But this is not going to be the sort of firm a small manufacturer like Open Connect can really talk to, suggests Brown, since it sells other people?s kit as well. While you are still in the process of building a presence, that is an eventuality you simply cannot allow for.
Knowing the market is one thing. Knowing the product set is another entirely, because the goalposts keep moving. One area in which these goalposts have not so much moved as upped sticks and wandered off somewhere else altogether is in intranets. The main problem facing anyone wanting to build one of these beasts and hang it off an older system is where to put the Web server.
It is possible to create a logical server in the software of the older system, and hang the whole thing off a Vax or mainframe, but it?s awkward and expensive ? two good reasons for pushing the Unix/NT box attached to the mainframe as the most effective option. It is also worth pushing these no matter what customers believe they want. What they really, really want is something that works, and not too expensively.
For this reason it can be worth trying to wean them away from their older systems, regardless of how painful they think this is going to be. Charles Malir, managing director of Web connection specialist Orbis, believes the whole mindset of customers with legacy systems is different from those people who genuinely want to buy something new.
?As far as we can see, people tend to leave their legacy system alone. They don?t want to mess around with it.?
He suggests that the skillset is diminishing as the years pass (see box on the year 2000 problem for the obvious fashionable objection to that statement) and says this means there is an increasing incentive for people to push their customers away from the older technology. ?Customers will replace the older boxes when they are brave enough,? says Malir.
He believes that even vendors will find their own skills less suited to supporting their older kit, and at that stage everyone will be pulling in the direction of intranets. Although this is the view of a Web connection specialist, it needn?t be any less valid for that.
It seems there?s a difference in what people can mean when they talk about legacy systems and legacy architectures. Certainly, there are means of exploiting legacy systems themselves in much the same manner as there is a way of selling more or less anything. But what requires a little more skill is linking them to new technologies and being able to advise on which bits need replacing and which bits don?t. This is where the earlier example of the IBM AS/400e comes into its own, as do recent announcements from Digital and a number of other companies.
Another thing worth noting is that not all legacy systems are as out of date as some of the hype might have you believe. Once again the AS/400 springs to mind. Only a few months ago, in February to be precise, IBM put out a cheesy press release which made it clear that a number of users felt so strongly about their AS/400 systems that they sent the company a Valentine?s card all about it.
In terms of yuk factor and sincerity rating that probably goes off the scale at either end, but it demonstrates that what is hyped as a legacy system and regarded as outmoded by the Java Beanies isn?t necessarily as dated as all that in the real world. Many accounting companies and departments still regard AS/400 as their workhorse system of choice, for the baffling and arcane reason that it actually does the job rather well, and hang the trendiness or otherwise of the system.
Either way, there must be an opportunity in there for the switched-on dealer. As Jones points out, if you?re an organisation with 60,000 users wanting to upgrade to version 4 of a particular system, then it?s much easier if you can do it from a central workstation with a few clicks of a mouse. This is what will happen with network computers when they arrive en masse, although Jones is the first to admit they are not there yet: ?The market is starting to pick up, but most of what we?ve done so far are pilots and demonstrations. We?re looking at really far ahead technology.?
At a recent briefing, consultant Dennis Keeling said most of the banks were looking for a means of unifying all their data and getting at it with a handful of clicks. He added that all of this data was currently held in disparate forms across disparate networks with different protocols that had been built up department by department, many of which were reluctant to let their pet technology go. The technical environments are all mixed, and it is unlikely that this mix will be the same as you move from site to site.
This in turn means that every legacy upgrade must by definition be different from the last, so if it?s an area that you?re keen to get into, it?s probably not a bad idea to treat the solution in a box merchants with extreme caution.
Learning to work with tailor-made systems is a must, but even after mastering this, it is necessary to take on another idea ? the old chestnut of business process re-engineering.
The thing is, if you?re introducing a more collaborative technology, you?re bound to end up changing the way people work. Empire-builders don?t like democratisation of data, they?d much rather hang on to their little pockets of power. There are dealers, resellers, Vars, call them what you will, that regard that side of the business as a bit above and beyond the call of duty. But it?s as well to bear in mind that if you?re determined to sell into a legacy environment, that it?s the sort of sale you could easily be looking at.
A final thought, though, and a sobering one, comes from Malir. He is not convinced that it is actually possible for resellers to get deeply involved in legacy systems, simply because of the likely profile of the customer and original vendor. ?The vendors tend to treat it as a bit of a cash cow,? he comments. ?They generally have a very high maintenance requirement and customers are tied in very closely to the vendor.?
In terms of the customers themselves, he points out that they are likely to be the larger corporate accounts. ?It?s difficult to imagine a smaller or newer reseller getting involved outside of those companies that provide third-party spares,? he says.
And supposing someone comes to you and says they have these 6,000 workstations that need upgrading and they are all hanging off an SNA network on a mainframe. Unless you are a large reseller already in that area, is it really the sort of business you want?