2018 is shaping up to be the Year of Open Source as industry giants scramble to invest in the software and infrastructure powering modern applications. This year has already seen Microsoft acquire Github and donate many of its patents to the Open Invention Network and now IBM has announced the biggest pure software acquisition deal in history in acquiring Red Hat. What does this mean for Open Source in general and what should we as Software Asset Managers be mindful of?
IBM is a highly-acquisitive company that’s cash-rich so it should come as no surprise that they’ve pulled off this gigantic deal. At a stroke it makes them the owner of much of the underlying tooling powering cloud computing. Red Hat is strong in hybrid cloud and responsible along with SUSE for pioneering the use of Linux in enterprises. Alongside this over the past decade they’ve been eroding Oracle & IBM’s market dominance in enterprise Unix. Solaris and AIX, once the powerhouse of enterprise datacenters, are little more than maintenance cash cows for their respective owners. Still big products and important to the bottom line but hardly on the radar of IT strategic decision makers as a go-to platform.
Software for IBM is all about the maintenance revenue, a market that we’ve reported on previously as being hugely inefficient and a rich seam of potential savings for Software Asset Managers. Anyone who has transacted with IBM for new software will know that of all the big vendors they offer the largest discounts. Build the barn and then milk the cows.
The maintenance revenue part of the Red Hat acquisition therefore makes perfect sense, but it isn’t the driver for the acquisition. It’s easy to think of Red Hat as being “Oh, that company that sells maintenance contracts for Linux” but that isn’t what this acquisition is about. It’s about grabbing control of the technology used to power hybrid clouds. Key products here are Open Shift (Platform as a Service) and Red Hat’s packaging of Open Stack (Infrastructure as a Service). That sector of Red Hat’s business is growing at 100% per annum and accounts for around 13% of revenue.
IBM have dabbled in cloud for a number of years via their acquisition of SoftCloud but have gained even less traction than Oracle. What I feel they’re acknowledging here is that they’ll never directly compete with AWS or Azure in providing the application layer for net-new customers. Like Oracle they’ll target hybrid cloud revenue by leveraging their on-premises presence with large customers. This hedges against declining on-premises perpetual maintenance revenue streams by providing a path to the Cloud that doesn’t involve a lift-and-shift to non-IBM technologies.
IBM have been somewhat of a white knight in copyright and patent battles sparked by the Open Source Software (OSS) movement. They’ve been a long-term contributor to Linux and first invested in Red Hat in 1999. Many IBM software products made their Linux debut on Red Hat first. IBM Global Services have been selling IBM Software-on-Red-Hat solutions for most of this decade.
There is however already much concern in the community, one that has been battered in the past by the Oracle purchase of Sun/Java and the SCO-Linux wars. SCO are still trying to sue IBM for copyright infringement some 15 years on. This concern comes from the appropriation of community works for commercial gain and it feels like we may be on the cusp of another battle with the recent announcement of license changes impacting cloud solution providers.
The challenge is that the OSS community is founded partly on the idea of taking on proprietary software vendors like IBM & Microsoft. Developers willing to contribute to a community project may be less willing to do so if it benefits a large corporation. Having said that, Red Hat themselves are a large corporation – almost $3bn revenue in 2017 – so that concern may be moot. Microsoft, Google, Red Hat, IBM, and Intel already dominate commits to open source projects.
IBM’s track record of assimilating innovative technology companies is patchy. Informix, Cognos, and SPSS have delivered good revenue from maintenance contracts but largely missed opportunities in the rapid growth of BI. Like AIX, they no longer have mindshare with decision-makers who instead plump for back-ends from database manufacturers and presentation layers from Tableau or Qlik.
These acquisitions though were small-fry compared to Red Hat.
As a long-term IBM-watcher the Red Hat acquisition sparks memories of IBM purchasing Lotus in 1995. A colossal price was paid, it made the front page of the Financial Times & Wall Street Journal rather than Computer Weekly, and it moved IBM in to a new-for-them environment – the desktop.
When IBM purchased Lotus it originally operated them as a separate business unit. The same is planned for Red Hat. Within five years however it had begun assimilating the business through centralisation of marketing and management functions. These changes coincided with the decline of the key acquired product, Lotus Notes. A number of poor marketing and product design decisions accelerated the decline and whilst Notes provides strong maintenance revenue it has been left to wither on the vine. For example, there was an unseemly scramble to patch the software to cope with recent changes to TLS/SSL support. Only recently has it been refreshed with development and product management outsourced to HCL India. The same is true for other IBM acquisitions Tivoli & Rational.
Clearly, IBM cannot allow this to happen to Red Hat, but the potential corporate culture clash is similar to IBM-Lotus. Whilst the stereotype of white-shirted, blue-suited IBMers is no longer relevant IBM have retreated from home-working & remote location staff deployments to centralisation in major hubs. IBM’s infamous HR practices also don’t fit Red Hat culture, and Ginni Rometty doesn’t engender the same amount of zeal and commitment from her employees as Jim Whitehurst does at Red Hat. Jim has a remarkable 94% approval rating on Glassdoor versus Ginni’s 67%.
Red Hat will face a battle to retain independence, but they might just be the catalyst for changing IBM’s culture. If Red Hat continue to grow as they have for the past 14 quarters (average growth 17% per quarter) they’ll be IBM’s best performing business unit for the foreseeable future. Success defines culture as much as culture defines success. With Ginni due to retire Jim has to be in pole position to assume her role. He has the right background and track record and is also the right age. Rometty’s tenure has been characterised by declining growth, cash-funded acquisitions, and widespread layoffs. With the purchase of Red Hat, she may just have cemented her legacy and made IBM relevant for the next decade. Jim’s people-focused management style would revolutionise IBM culture.
With any acquisition comes a degree of uncertainty for Software Asset Managers too. Red Hat’s existing software licensing agreements are summarised in our article here and they already have sufficient complexity to result in license compliance risk, particularly in the areas of Disaster Recovery and Matching Support levels. Red Hat acknowledge this complexity and provide compliance tooling as part of their Satellite add-on subscription. Think of this as similar to SAP LAW. Other tool vendors including Snow have Red Hat-specific modules and reporting.
Red Hat are assertive compliance auditors and in this regard they are meeting fellow travellers and familiar business processes in this acquisition. One area of potential difference, and therefore change, is that Red Hat don’t use performance-based metrics like PVU or RVU. Most Red Hat software is licensed based on socket counts without regard to the power deployed from those sockets. If IBM wanted to extract more revenue from the Red Hat acquisition a move to PVU metrics is an obvious shortcut.
The final question here is who’s next? Will Ubuntu-publisher Canonical court an investor? Do they need to? By way of comparison fellow enterprise Linux publisher SUSE was recently sold by Micro Focus for $2.5bn. Canonical have rarely turned a profit which points to potential economies of scale as being a way forward for them. They post similar growth numbers to SUSE & Red Hat, yet SUSE made a profit of $56m in 2017 on revenue equal to Canonical’s. Clearly there are efficiencies to be found there. Canonical is another particularly strong vendor of the OpenStack and hybrid cloud services that attracted IBM to Red Hat.