Open Enterprise

RSSSubscribe to this blog
About Author

Glyn Moody's look at all levels of the enterprise open source stack. The blog will look at the organisations that are embracing open source, old and new alike (start-ups welcome), and the communities of users and developers that have formed around them (or not, as the case may be).

Contact Author

Email Glyn

Twitter Profile

Linked-in Profile

If Oracle Bought Every Open Source Company...

Article comments

Recently, there was an interesting rumour circulating that Oracle had a war chest of some $70 billion, and was going on an acquisition spree.

Despite the huge figure, it had a certain plausibility, because Oracle is a highly successful company with deep pockets and an aggressive management. The rumour was soon denied, but it got me wondering: supposing Oracle decided to spend, if not $70 billion, say $10 billion in an efficient way: how might it do that? And it occurred to me that one rather dramatic use of that money would be to buy up the leading open source companies – all of them.

That would be eminently possible, since the largest such company - Red Hat – is worth around $6.25 billion; the other candidates would be fractions of that sum. And since Oracle would only need to buy one company per sector – one for enterprise content management (ECM), one for CRM, one for business information (BI) etc. - and already owns MySQL, the total required for a complete "set" would be relatively modest. It could certainly afford to do it. But why might it want to take this course of action?

Offering extremely low-cost ancillary programs like ECM, CRM and BI would allow it to enter new markets which are currently unable to afford the high-end Oracle database application. It could do that by offering MySQL as a kind of “taster”: it wouldn't be losing any revenue, since these companies wouldn't have bought the main Oracle database offering anyway. And once they are happy using MySQL and other Oracle open source products, some might eventually be persuaded to migrate to the “full” database application.

There would be other benefits, too. It would give Oracle control over a wide range of applications that it could use against its main rivals. Even if it did not make much money from these open source offerings, it could reduce the amount others made from their high-cost proprietary applications. That, in turn, would reduce their scope for competitive action against Oracle.

Owning all the main open source programs would give Oracle increased influence with other players in the computing world – for example, hardware manufacturers who might want to bundle apps, system integrators etc. Rebranding the open source programs as part of the Oracle family, and then encouraging their wide distribution as free downloads would provide valuable marketing and boost the Oracle brand – not least in developing countries unable to afford the main Oracle enterprise offerings.

It's clear, then, that many advantages would accrue to Oracle from swooping in and buying up all the leading open source companies. But even a cursory glance shows that it doesn't look such good news for the free software world.

First, it would effectively replace a dynamic ecosystem of many smaller players with one dominant entity. Worse, that owner would be aiming to use open source for its own benefit, rather than nurturing the open source ecosystem as a whole.

It would allow Oracle to dictate terms to practically any company involved in the open source business world. It could unilaterally set standards – including non-open ones – in the open source domain. It would mean that it employed a large proportion of all open source coders that were paid to work on free software. Clearly, this would be a monoculture just as bad – albeit very different – from the Microsoft one that open source was supposed to offer an alternative to.

So, let's look at what prevents such a cataclysmic development from taking place. Sadly, there's little that the companies themselves could do. The ones that are publicly traded could hardly resist, and even those that remain privately held would find their backers quite happy to accept large wads of foldables from Oracle: after all, cashing in is precisely why venture capitalists invest in startups in the first place, and if Oracle made them an offer they couldn't refuse, they wouldn't.

If the companies and their management couldn't do much, what about the coders? Here, things are more hopeful. If large numbers of hackers simply upped and left, the value of the open source portfolio would be hugely diminished. Rather foolishly, Oracle has already shown with its acquisition of Sun and its collection of open source projects that it either doesn't know how to manage such coding talent – or really doesn't care.

Either way, employees of open source companies that were acquired by Oracle would know that their prospects were not particularly bright, and that they would not be furthering the cause of free software much by staying in its employ.

That means many of them would probably leave, either immediately, or in due course (depending probably on how their share options worked.) Similarly, there would be managers and investors looking to fill the vacuum left by Oracle's sweeping up of the main players in the open source business world. That's clearly a great combination for creating new offerings in the relevant sectors. An additional impetus would be the fact that users of those applications might also be open to moving to new programs that weren't all tightly controlled by Oracle.

The question would then be, should such companies fork the old open source applications, or start from scratch? Forking would have the advantage of familiarity for coders and customers alike, and it would allow current users to migrate more easily. On the downside, new startups would be obliged to contribute all their code additions back to Oracle, which could then offer it under a proprietary licence unable to offer a proprietary licence for the forked code, because of the terms of the open source licence it was released under. This would place new players offering the forked version at a disadvantage.

It might, therefore, be better to create an entirely new code base, which would create a level playing field. It's true that this would place an obstacle in the way of migrations, but it's not as if users are locked in to open source programs in the same manner they tend to be with proprietary ones.

However, it would be important not create a situation that would allow a repetition of Oracle's wholesale purchase of an ecosystem. To prevent that happening again, it is important to choose the right model for licensing the underlying code. There are two main options here that would prevent the kind of issues that a mega-acquisition on the part of Oracle – or anyone else – might create.

These really arise from the fact that the copyright holder of a program licensed under the GNU GPL – still the most popular licence for open source companies - has a privileged position in that it can also license the code under conventional commercial terms. Those who fork the code can still make money from it, but they are not able to offer such proprietary versions.

Perhaps the best way to ensure parity for all players is to assign all the copyrights to an independent foundation – just as the Free Software Foundation owns all the copyrights to programs in the GNU project. This ensures that no company has a privileged position: they must all offer code under the GPL, or AGPL in the case of a cloud-based solution, and make any changes that they distribute available to all.

Alternatively, companies might choose to release their code under the Apache licence. This takes the opposite approach to creating a level playing field by allowing *anyone* to offer versions of the code under a commercial licence, not just the copyright holder. Clearly, that is less satisfactory from the point of view of preserving users' freedom, and so which approach is chosen is probably down to the personal preferences and beliefs of the people setting up the new companies.

The point of this thought experiment is not so much to suggest that the open source community needs to be taking active measures to prevent it being assimilated by Oracle (or any other large player), so much as to encourage some serious thinking around the perennial issues of licensing. This is not some marginal area of interest only to geeky lawyers (or legalistic geeks), but goes to the heart of what an open source company is and does. In addition, as the above has indicated, it is a major factor in determining not just the fate of the company in question, but also of the wider open source ecosystem.

Follow me @glynmoody on Twitter or

Email this to a friend

* indicates mandatory field

ComputerWorldUK Webcast