Comment Not too long ago, selling middleware and tools to software developers was a big business.
Large technology empires were built on a single premise: that computers need software, written by software developers who need a panoply of infrastructure tools and middleware - and that you could charge the developers for that infrastructure. Household names like Borland, BEA, and Rational chugged along merrily.
More ReadingNokia tries to lure Android porters with free Nokia X mobesPivotal: So who fancies skinny-dipping in our 'Business Data Lake'? PS: It'll cost youSatya Nadella shakes up Microsoft, appoints 'Scroogled' man Mark Penn as strategy chiefHey, IT department! Sick of vendor shaftings? Why not DO IT, yourselfJava or .NET bod in the Midlands? Congrats - you've got a DOUBLE DIGIT payrise
While vendors counted their middleware lucre, developers debated "Java vs .Net" like it was some kind of important existential concern. Disruption came from marrying up the rainbow-and-sandles world of FLOSS to the free-loading spirit of the web... and the developer market quickly had its Napster moment. Free (and - sure, sure - open source for those who care) technologies like Perl, PHP, Python, and Ruby coupled with free, "enterprise grade" web services like Apache (and then nginx) all but eviscerated the "developer market" in the 2000s.
The finishing slice came when the Java ecosystem open-sourced, through JBoss, and then Sun capitulated to open source, and finally the perfection of SpringSource arrived. The list goes on - MySQL, anyone? Soon, the idea that you'd make money by selling to developers was laughable. Instead you made money by selling the company to another company: MySQL sold to Sun for $1bn, and SpringSource sold to VMware for around $400m.
You've got to make friends
While the developer market withered, developers remained an important force, "king-makers" even, as my former colleagues at RedMonk put it. Winning developer allegiance is still key to building ecosystems around companies and platforms from historical cases like Microsoft and Java, to Apple's success in recent years. But making money from developers directly was another story.
A developer might shell out $50 to $100 for a nice text editor or IDE: but pay for middleware? Nope. Running that middleware as a cloud-based service, however, is a clever hack around this parsimony. Developers are driving much of the spend on IaaS and PaaS.
Developers need somewhere to run their software, and buying new servers doesn't appear to be where they're putting many of the new applications. There's certainly a market there: taking out SaaS, [our bottoms-up forecast at 451] puts the public cloud market at around $20bn in 2016, with the majority in IaaS and about 25 per cent in PaaS. When you look at Amazon's portfolio, it's essentially a stack of middleware, all run, and charged "as-a-Service."
A quick look at the customer logo-porn pages of competing IaaS and PaaS providers shows the importance of developers; many of these companies, like Rackspace, have established developer relations programs and before getting snatched up by IBM, SoftLayer was self-admittedly focused almost exclusively on developers. With all the hand-wringing about "shadow IT" - namely developers going out-of-band to pay for cloud-resources - the connection between developers and spending is even more direct. It seems developers are perfectly happy to pay for middleware, if only you'll run it for them.
Follow the money... the VCs' money
Incubators like HeavyBit are betting on the return of the developer market as well. Companies in their portfolio service developer needs, from IDEs like Codeenvy, to QA platforms like Rainforest, to the API documentation service Apiary. Of course tracking VCs' bets is more about doing once-removed crystal-ball-gazing on what the mega-vendors would acquire in three to five years than accurately tracking broader trends.
And to that point, IBM's movement back towards the developer market at its recent Pulse conference provides one of the better indicators for which direction the developer market is going. Over the past year, IBM has been reorienting large swaths of the software group's portfolio around not just cloud, but cloud as a medium to go after developers.
Armonk's finest tripped over each other to paint the vision of enterprises getting back into software development, driven by the need to interact with customers in new and exciting ways ("social") and on new and exciting platforms ("mobile," as well as tablets).
Big Blue released a beta of its new middleware stack, BlueMix, based on Cloud Foundry. Filling out the new middleware stack more, IBM picked up Cloudant, a Database-as-a-Service company, one of those previously free chunks of downloadable middleware that's now Monetisable-as-a-Service now. Indeed, at 451, we estimate the DBaaS market was worth $150m in 2012, growing at a CAGR of 86 per cent a year to $1.79bn in 2016, attractive to any vendor strategist.
With all of the cuts and bottom-line optimisation going on at IBM, this much focus on an entirely new middleware stack, targeted at developers, is a large signal of intention.
IBM's not the only one placing developer bets. VMware and EMC bet on developers in 2013 when it combined the SpringSource, Cloud Foundry, and Big Data assets together into Pivotal, with ex-VMware, now Pivotal CEO Paul Maritz saying they were building "a new platform for a new era." This was quickly followed by GE $105m investment into the developer-oriented company.
These "new platforms" are certainly out there, from the Jawbone Up that tells me I don't walk enough and sleep poorly, to new marketing applications that better track and seek to program my buying habits. Where there's software, there are developers.
Will the developers finally pay for the tools they use to make their write and run software? In the consumer space of $19bn exits, oddly enough, perhaps not: many of the old ways hold true – there is still DIY pride and 20-year-olds with nothing better to do than code all night.
Outside of the Ramen-noodle-coated technology world, however, as more devices get IP addresses and need software accordingly, it's not full-on bonkers to think that there will be more developers at "normal" companies. And that's the meat-and-potatoes of any "infrastructure" play: the mainstream companies which would rather purchase tools and middleware than quickly polish off another cup of Ramen before firing up a bare-bones editor to type up yet another chunk of middleware from scratch. ®