[I did actually get an offer to buy this piece, which is a few months old now, but I was just changing publishers and decided it would be impolitic. So, have a freebie. :¬) ]

After the on-again off-again IBM deal, Sun Microsystems has agreed to sell itself to Oracle for £5 billion – a deal which both companies had managed to keep very quiet.

Unarguably, there is vastly less overlap between Sun and Oracle than with IBM. If the earlier deal had gone ahead, there would have been rivalry between SPARC and POWER, between Solaris and AIX, between two server ranges, two storage ranges and much else besides. But there will still be blood on the floor in the Oracle merger. Oracle has its own enterprise operating system, Oracle Unbreakable Linux – essentially a rebadged version of CentOS, which is itself Red Hat Enterprise Linux with the names and serial numbers filed off. OUL has not been a huge success – not that OpenSolaris has, either – but Solaris still has considerable traction and goodwill attached to it among the sort of high-end enterprise Unix people that still don't entirely trust Linux. Sun's flirtations with Linux have long seemed a little uneasy, and the Oracle deal will probably end them.

Otherwise, though, it's a good fit. Both companies are big-iron sort of vendors, selling mainly to large enterprises. Both favour Java. Many Oracle installations run on Sun kit – although the deal will certainly make life interesting for salespeople trying to push Oracle systems on other vendors' hardware. Still, an all-in-one hardware-to-software-to-integration-consultancy sell may prove distinctly advantageous up against major rivals HP and IBM.

A significant issue could be over the two's open-source database programs. Sun owns MySQL, the leading FOSS database which powers a million small websites and blogs, as well as a few big ones. Oracle owns Innobase, one of the leading storage engines for MySQL. The whole ensemble being under one of the biggest of big corporate roofs will not reassure FOSS enthusiasts. A careful VMware-style spin-out here might appease the ravening GNU hordes (or should that be herds?)

Sun's SPARC server line may also be a difficult sell in an increasingly x86-dominated world. Fujitsu are a major partner here and a sell-off might tempt the newly-merged company, but it's far from certain Fujitsu would be big enough to go it alone with SPARC – if it were interested at all.

A lot of people in the IT industry won't be interested at all. Today, an entire generation of IT staffers, from helpdesk minions to IT directors, know nothing but PCs and Microsoft. For them, the answer is Microsoft, now what's the question; the only one they ask is if anyone is cheaper than Dell for the commodity boxes to run it on. It's an increasingly hard sell to persuade such types that for big systems, non-Windows systems can offer much better scalability, continuity, performance and uptime. When you're used to rolling out hundreds of little Exchange boxes, each supporting hundreds of people, the notion that one big – and expensive – box could support all of those tens of thousands is scary and alien. It's getting even harder now that virtualisation is legitimising the argument for ever-bigger piles of commodity, “industry-standard” kit. The result is lots of little virtual Windows servers running on a few honking great Windows servers, and a veritable nightmare of patching, updates and zero-day exploits rampaging across an IT monoculture built around tools originally designed for desktop PCs, with security, scalability and manageability bolted on later.

But that is the pitch that Sun/Oracle's sales lizards will have to make. Small companies grow from little ones, and little ones start off using tools well-suited to little, cash-poor businesses: cheap generic PCs, cheap generic software and cheap generic IT staff. Sun/Oracle's big systems are arguably a much better fit for big corporations, but getting to there from here is a scary and expensive journey. The world depression will thin out a lot of the big old dead wood from the corporate forest, though, and the fast-growing saplings rising to replace them could be a lucrative market.

The new company will be a giant to rival HP and IBM, but much less inclined to hedge its bets than these two. IBM is dissociating itself from the commodity PC market, focussing on consultancy and its own proprietary large server systems and software, bridged to the wider world via an open-armed but vendor-agnostic approach to open source and Linux. HP is doing quite well being the big daddy of serious corporate PC kit, from pocket devices via high-volume workstations to big servers, while de-emphasising its own legacy proprietary boxes running Itanium, HP-UX and OpenVMS. It increasingly just leaves the software side up to Microsoft.

Sun/Oracle has a bolder, more difficult road to tread: its own servers and storage and its own bottom-to-top proprietary software stack – albeit judiciously open-sourced or based on industry standards in places. It's a good stack, but the new company will boldly stand alone. Let's hope it doesn't end up getting hobbled, to fall and have its congress of clever technology dismembered by the less-proud pragmatic we'll-get-on-with-our-enemies approaches of HP or IBM.

Despite a noble history of powerful workstations and some industry-leading design and build quality, Sun/Oracle barely has a presence in the desktop market any more outside of SunRay thin clients, and its desktop software has not exactly received an orgasmic welcome. There is another acquisition candidate here, who lead the world in immaculately-designed hardware and proprietary software, but whose entry-level servers are languishing. Currently, it's doing just fine, but if its charismatic but ailing founder and CEO "does a McNeally" and hands over the reins, Apple would fit into the new Sun/Oracle empire quite nicely.