How Red Hat can catch the developer train

Developers love open source. So why don't they love the open source leader?

Ask a CIO her choice to run mission-critical workloads, and her answer is a near immediate "Red Hat." Ask her developers what they prefer, however, and it's Ubuntu. That's just for the operating system -- Red Hat's strongest product franchise.

Outside the operating system, according to AngelList data compiled by Leo Polovets, these developers go with MySQL, MongoDB, or PostgreSQL for their database; Chef or Puppet for configuration; and ElasticSearch or Solr for search. None of this technology is developed by Red Hat.

Yet all of this technology is what the next generation of developers is using to build modern applications.

Given that developers are the new kingmakers, Red Hat needs to get out in front of the developer freight train if it wants to remain relevant for the next 20 years, much less the next two.

What developers want
Developers, writing applications for lines of business within enterprises large and small, increasingly deploy to the cloud. Whether looking at the leading public cloud (Amazon Web Services) or OpenStack, developers have been voting overwhelmingly for Ubuntu, not Red Hat.

According to the Cloud Market, which tracks adoption on AWS EC2, Ubuntu claims more than 50 percent of all guest OSes. Across Red Hat's complete product range (RHEL, CentOS, and Fedora), the company gets less than 20 percent of all guest OSes. On OpenStack, where Red Hat has been the top contributor for the last few releases, Red Hat notches a more respectable 34 percent adoption rate.

Ubuntu? 55 percent.

Red Hat isn't merely an operating system company, of course. It has a thriving middleware business (JBoss), a growing storage business (Gluster, Ceph), and an increasingly strong footprint with developers through its OpenShift and OpenStack cloud products.

Still, for a company that pays the bills by making it easy for mainstream enterprises to consume otherwise unwieldy open source software, Red Hat has missed out on nearly all of the biggest projects, from Hadoop to ElasticSearch to Cassandra. Each of these technologies is now backed by private companies that are valued too richly for Red Hat to buy them, but Red Hat also lacks meaningful partnerships with most of them.

These are the technologies that developers crave, as Polovets' analysis shows, yet Red Hat offers very, very few of them in any meaningful way.

What open source companies want
This isn't to suggest Red Hat offers no value. Clearly, it offers a great deal, translating into more than $1 billion in annual revenue. The key is who pays that money to Red Hat: operations.

Way back in 2003, Red Hat made the very difficult decision to focus completely on operations professionals and largely to eschew developers. By introducing Red Hat Enterprise Linux (then "Advanced Server") and tying its brand completely to a commercial, enterprise-grade product, Red Hat was forced to downplay the more developer-friendly Fedora.

The move made financial sense. But it didn't make developer sense.

Most open source companies have taken the opposite tactic, starting with developers. SpringSource, MySQL, JBoss, and others fit into this category. What JBoss taught us is that dev is critical to fueling adoption, but ops is critical for paying the bills. Yet no one besides Red Hat has managed to effectively capture those operations budgets at any sizable scale.

The best route to those budgets may be Red Hat.

What Red Hat wants
Red Hat needs developers. It also needs a big story consistent with its brand as the leading open source company. Meanwhile, the most promising open source big data companies have developers, but now need a stronger story with operations as they seek to turn developer affection into access to operations budgets.

Can you sense the symbiosis?

Developers are flocking to Hadoop, NoSQL technologies, configuration management systems, and Web frameworks, among other items. Virtually all of the best technologies in these categories are open source.

One way for Red Hat to become more relevant with the rising generation of developers is to more firmly tie itself to the companies and projects that these developers love, either through acquisition or, more likely, through partnerships.

On the acquisition side, Red Hat has made bold, strategic acquistions in the cloud space (Ceph, KVM), but has not made a significant, developer-oriented acquisition since JBoss. Imagine the developer love that would flow to an ElasticSearch/Kibana/Logstash combination.

More likely, given the valuations for developer-rich companies, is a partnership. Here Red Hat has been active, to a degree. Red Hat, for example, has been strengthening its relationship with Hortonworks, building out joint solutions and go-to-market strategies.

But this relationship omits the Holy Grail: distribution through Red Hat Enterprise Linux and a product SKU that Red Hat's sales force will push directly and through its channel.

The problem is that Red Hat, like any good company, must focus, and its sales force, not surprisingly, has never been adept at selling others' technology, given that it has its hands full managing the burgeoning portfolio of Red Hat technologies, including everything from JBoss middleware to virtualization technologies (RHEV) and now cloud.

What may be reasonable middle ground
It may be a bridge too far to expect Red Hat to sell others' technology. But it shouldn't be too much to expect Red Hat to help distribute it. The question is how, because getting into a co-marketing arrangement with RHEL is very difficult.

Sometimes it happens and the result is spectacular. Take Docker, for example. Docker is one area where Red Hat discerned developer interest and partnered early and deep both with the company and the technology. Red Hat announced a container certification program around the last Red Hat Summit in March and OpenShift has used Docker extensively.

Speaking of OpenShift, this is another example of a developer-centric Red Hat offering. While OpenShift started slowly, I've heard from a number of sources that OpenShift adoption is booming now, fed by a steady stream of developers.

All this is good and shows that Red Hat recognizes where it's weak and is taking steps to remedy the problem. But more is needed, and that more probably comes back to its primary platform: Linux.

In the past, Red Hat tried to find ways to promote other open source solutions without actually selling them itself, Red Hat Exchange being a prominent, if unsuccessful, example. More recently, Red Hat introduced Software Collections as a way to loosely couple fast-moving developer favorites like Ruby and MySQL with slower-moving enterprise favorites like RHEL.

Red Hat's community operating system
It's a good start, but it still treats these developer technologies like arms' length, second-class Red Hat citizens (somewhat akin to Microsoft embracing open source but only through a separate subsidiary, Microsoft Open Technologies). As I noted before, getting into RHEL is brutally hard.

It might also be unnecessary --at first.

A better strategy might be to create a truly first class developer-oriented big data stack on CentOS. CentOS has the ability to offer both compatibility with the RHEL ecosystem and to provide more developer-friendly stacks on top, delivered through CentOS Special Interest Groups.

Another great Red Hat platform is Project Atomic, a new community project to develop technologies for creating lightweight Linux container hosts. As described by Red Hat engineer Colin Walters, Atomic "could evolve the application from something that is installed to something that is communicated," which, in turn, "could drive a further, and much needed, decomposition of the operating system layer into something lighter-weight, easier to manage, and less restrictive."

Atomic, then, has real potential to steal adoption from Ubuntu in AWS if done right.

Community now, money later
Whether we're talking CentOS or Atomic, however, it's clear that Red Hat's developer outreach probably doesn't begin with RHEL, even if it ends up there. Developers want to get productive, fast. Forcing them into a commercial relationship is the wrong way to achieve this. Such developer technologies need to be exceptional in their own right, not a thinly veiled appetizer for "real" products like RHEL or RHEV.

As Red Hat embraces and develops community-focused technologies like Atomic, CentOS, and OpenShift, then works with leading, complementary open source projects to ensure they get first-class treatment on Red Hat's platforms, Red Hat will appeal to the developers that crown technology winners even as it provides a clear path for that developer love to translate into Operations cash.

This helps Red Hat by making it relevant to developers again. It also helps such projects by giving them access to Red Hat's distribution. Equally important, the more Red Hat and these projects and companies work together on developer outreach, the easier it becomes to take the next step and create commercial strategies together.

This story, "How Red Hat can catch the developer train" was originally published by InfoWorld.

Join the discussion
Be the first to comment on this article. Our Commenting Policies
See more