Strange bedfellows

Following the buyout of Tandem by Compaq, industry watchdogs have been left wondering whether it will prove a mighty union or a huge mistake. Only time will tell, says Sean Hallahan.

Never has the old phrase chalk and cheese been more appropriate thanen left wondering whether it will prove a mighty union or a huge mistake. Only time will tell, says Sean Hallahan. when Compaq took over Tandem earlier this year. It was certainly good news for both companies, but bad news, it seems, for just about everyone else.

Both Tandem and Compaq, in their respective fields, have been innovative and aggressive. Historically, they have upstaged their major competitor, IBM, and forced other rivals on to the back foot. But despite their strengths - Compaq as the leading PC supplier and Tandem as the premier manufacturer of enterprise fault-tolerant hardware and software - both have suffered from weaknesses in the past.

In Compaq's case, its products effectively stop at the desktop. Even though it plays an important part in the PC-based server market, it is still viewed as a PC supplier, with no mid-range offering to challenge the likes of Hewlett Packard, Digital and IBM's AS/400 and RS/6000 machines.

IN AT THE HIGH END

For Tandem, the trouble is that it is viewed purely as a data centre box, running high-end online transaction processing systems, but scarcely having a presence outside that arena. It made forays into the Unix arena and struck up deals with PC suppliers to offer a wider service, but the market never took it seriously, despite the fact that its Unix offerings were highly praised by independent analysts.

But not all analysts are convinced Compaq has made a wise decision in bidding for Tandem. In a report on the merger, US consultancy firm Forrester Research suggests that Compaq may phase out Tandem's Mips chip-based systems in favour of Intel and downgrade investment in Tandem technology.

Forrester stated that Compaq would have been better off picking up another NT and Intel-based mid-range to high-end manufacturer, for instance Digital or Data General. The report also suggests Tandem's direct sales force, which will now be selling the entire range of Compaq/Tandem products from the desktop to mainframe, will cause conflict within the channel.

A diametrically opposing view is taken by UK consultancy firm Bloor Research.

'It is a marriage made in heaven. Both Tandem and Compaq have been working on NT solutions and Tandem has already started to move over to the Intel chip,' says Robin Bloor, the firm's chairman and chief executive. One of the advantages of the merger is that there is effectively no product overlap between the two organisations.

Bloor rejects the idea that a fusion between Compaq and Digital would have made any sense. 'Compaq could have bought Digital, but it has competing products. In cases such as these, the products of the company being purchased would be canned. There would be a period of about a year when Compaq would go through Digital with an axe.'

Tandem employees, both past and present, are unwilling to go on the record about what they think the future will hold for the combined company. One former senior manager who worked at Tandem for nine years - and wants to remain anonymous - has clear views on the situation. 'It's a good buy from Compaq's point of view. From a Tandem point of view, I think they were bloody lucky,' he says.

An existing senior employee, who heads one of Tandem's divisions, was also unwilling to be quoted, but says: 'My view is that it's a good deal for both companies. All the reports have said the acquisition extends Compaq's range from the desktop to the data centre, but it does much more - it extends its technology base and gives it ownership of ServerNet.'

ServerNet is a product designed to provide 'sophisticated, non-saturating communication capabilities and processor-to-processor communications at 300Mbps per router,' states Bloor Research's report Parallel Database Technology. 'To our knowledge, this is faster than any multistage switch technology from any massively parallel processor manufacture.' Industry wags have already been suggesting names for the new company. The front runner is a combination of the first syllable of the name Tandem and the last syllable of the name Compaq - Tampaq.

NICHE TO MEET YOU

One of the problems Tandem has faced in recent years is in its niche market - a non-stop, fault-tolerant, high or continuous availability mainly for the financial and defence community - which has come under attack from other vendors. For two decades, Tandem had an almost unique offering unmatched by any other supplier.

Even IBM, as it has done so often in the past, ignored the threat presented by Tandem technology. Tandem machines were often used as a front end to IBM mainframes in the financial sector. In 1985, over a decade after Tandem was founded, IBM was forced to recognise that it was missing out on an opportunity in the online transaction processing (OLTP) market, and so signed a deal with Stratus, another fault-tolerant supplier, to badge and sell its machine as the IBM System 88.

That setup continued until the mid 90s, when IBM mainframe technology was robust enough to offer near continuous availability. Tandem's domination of the non-stop, fault-tolerant market was at an end.

To some extent, Tandem recognised that its window of opportunity was closing. In 1993, it switched from Cisc to Risc architecture with the move to Mips and moved to an almost open systems position.

But the move was not as easy as Bloor's report, New Computer Hardware: Options and Comparisons, pointed out in 1995. 'The challenge for Tandem lies in the changes to marketing and relationships implied by this new tack.'

WOO AND VAR BETWEEN

These challenges essentially involve the replacement of former high margins with volume sales and the wooing of Vars and software houses.

The Unix market is application driven, and Tandem can no longer afford to be marginalised as a niche player - it must take success into the mainstream.

It has already had some success with Vars, but needs more. 'This is less a technical difficulty than one of perception,' the report states.

In the same report, Bloor Research considers the future of major manufacturers and the changes taking place in the past few years. Defining a major hardware supplier as a 'company that designs, manufactures and sells mainstream computer systems', it identifies only two long-term candidates: IBM and Hewlett Packard.

But the report goes on to pinpoint Compaq as a potential challenger to these two. 'A stalking horse could be Compaq if it can configure its PC engines into mainframe class systems and position itself credibly to sell this type of system,' it says. The merger with Tandem may just allow it become a world-class enterprise systems supplier, but it will still have some hurdles to overcome - technological, marketing and cultural.

There are significant cultural differences between the two companies in their approach to the channel. Compaq has always remained committed to the channel, while Tandem has largely relied on its direct salesforce to sell high-end systems.

If the takeover of Tandem by Compaq is viewed as something like salvation from decline for the non-stop manufacturer, then for the PC company it represents a half-open door to the data centre. Compaq has recognised that its continued growth depends on being able to move beyond the desktop server and desktop client into the data centre proper.

Due to the close policing of US takeovers by the country's watchdog, the Securities and Exchange Commission (SEC), Tandem was unwilling to comment on any details of how the two companies would meld together and what future plans they have, if any. 'There is a diversity of views, but there always is in these situations because people are coming at it from different positions,' says a Tandem representative.

PRESS EXPRESS

According to another representative, the views of Tandem's customers reflect the diversity of opinions expressed by the analysts and in the press, although he feels the prevailing attitude is positive.

Phil Payne, an analyst at Sievers Consulting, believes that Compaq had a two-fold motivation for acquiring Tandem - getting its hands on ServerNet and boosting its salesforce. 'Compaq already said it wanted to increase its salesforce by 3,000 - by acquiring Compaq, it has added 2,000 straight away. Compaq wants ServerNet to secure its own clustering technology and keep them out of the way of Microsoft's wolfpack,' Payne says. But he does not see the merger providing a threat to IBM's mainframe data centre business in the near future.

It will be some months before the intentions of the merged company become clear. The new company may not put much pressure on IBM, but if it works it will certainly offer a challenge to the likes of Hewlett Packard and Digital as well as others, and by implication to their resellers.

AND IN THE BIG BLUE CORNER

Both Tandem and Compaq have a history of challenging IBM and winning market share, sometimes directly and sometimes through an unofficial alliance with major rivals.

Both companies began life as niche players - Tandem in 1974 and Compaq in 1982. The original Tandem designers quickly spotted that the growth of networking, particularly in banking, required a message-switching system offering not only high availability but also non-stop processing. Its slogan upon UK launch in 1979 was: 'No single hardware or software fault can cause this system to fail.' The company went on to act as a front end processor for many of IBM's financial mainframes, eventually forcing IBM to sign an OEM deal with rival Stratus, and sell the machine as the IBM System 88.

Compaq's origins lie in the PC revolution of 1981 when IBM released its first machine. Compaq's offering arrived a few months later, but with one important difference - it was portable. Unlike many other clone suppliers, Compaq realised it needed a differentiator. Portability (although the two-stone machine can hardly be called portable) was it. It was not until 1984 that the company first challenged IBM on the desktop with the launch of Deskpro, which offered greater performance at a lower price. Two years later, Compaq again beat IBM to the punch with the launch of an Intel 80386 machine. Big Blue, anxious to maintain profitability on the existing 286 range, delayed seven months before following suit and ceded the technological leadership to Compaq.

Tandem, meanwhile, was capitalising on its niche market. Like Compaq it was innovative. One of its first moves was to produce a relational database, Non-Stop SQL, which was a success when equivalent products were being criticised for lack of performance. It also gave Tandem a chance to break out of its OLTP niche, although it subsequently failed to do so. Up until the end of the 1980s, Tandem was doing well in the OLTP environment, but things were changing. Unix was more mature, hardware less expensive than proprietary-based systems and more commercial packages were emerging.

All Tandem's application software was bespoke, written exclusively for a specific customer.

Compaq's forays into Unix, like Tandem's, were never highly successful.

If the late 1980s spelt trouble for Tandem, the early 1990s proved similarly difficult for Compaq. The company discovered that cheap capable machines made in the Far East were being mass-produced and marketed against its own and IBM's brand.

Compaq, which had made its name on performance and support rather than just price, found itself being undermined. The company's response was to cut prices. Tandem suddenly found IBM was able to offer similar fault-tolerant solutions to those which it had pioneered.

Both companies needed to find a partner - Compaq to take it beyond the desktop and Tandem to find a way out of a narrowing niche. With luck, one or two fortuitous disasters by competitors and a fair following wind, the new Tandem/Compaq alliance could be on to a winner.