By Charles King, Pund-IT, Inc. August 17, 2015
Linux saved IBM’s life in the 1990s. Though the company is in far different and better shape now than it was then, recent developments including the new LinuxONE solutions suggest IBM hoping to capture similarly dynamic benefits from its work with today’s open source community. Whether you agree with that or not, it’s worth considering the similarities between IBM’s first and the company’s latest Linux-driven strategies.
The little penguin that could…
Given its mainstream use in global data centers, it may be hard to parse the position of Linux in the mid- to late-90s. Initially released in 1991 as an alternative OS for PCs, by 1998 a sort of religious fervor had grown up around the platform with some folks claiming Linux could and would eventually unseat Microsoft’s market dominance.
Such beliefs may seem quaint today but at the time, with Microsoft mired in antitrust litigation and Apple stuck in what seemed to be endless Slough of Despond (Steve Jobs had returned to the company but its glory days were years ahead), Linux’s cute penguin mascot looked more formidable than it might have otherwise. There were more than a few practical problems standing in the way, however, the largest being the lack of a consumer-friendly GUI.
The open source collaborative development model made creating so complex a technology difficult, at best. However, Linux’s deep connections to and similarities with Unix – the same qualities that enhanced its appeal among technically astute professionals and hobbyists – helped make its case as a data center solution. Given its ability to run on multiple hardware platforms, including x86, and the swift evolution of Intel-based servers Linux pilot projects began in numerous IT organizations, often running on repurposed hardware.
The IBM factor
That said, Linux remained an intriguing “science project” for most commercial customers until 2000, when IBM became the first Tier-1 IT vendor to declare its support for the OS. That shocked many in the enterprise community, especially since the company declared that its recently rebranded zSeries mainframes would be the first target for new Linux applications and workloads. IBM’s vote of confidence provided Linux significant marketplace traction, as did the $1 billion the company committed in 2001 to expand Linux development.
But what did IBM get for its investments?
That can be divided into tactical and strategic issues/benefits. In the former case, though the IBM mainframe remained hugely successful in large enterprises, its adoption in new markets and among smaller organizations was problematic, largely due to steep upfront capital investment requirements. In addition, the success of client/server technologies had made those platforms, not IBM’s zSeries, the primary targets for many new applications and workloads. Finally, as the mainframe approached its 40th anniversary (in 2004), it was becoming increasingly difficult to attract young IT professionals and developers to the platform.
Supporting Linux and committing sizable funding to its development substantially addressed all of these points. In fact, the success of the Linux on zSeries effort led IBM to embrace and support numerous other open source projects, a practice that continues to this day.
Linux as an IBM strategic driver
What about the strategic benefits arising from IBM’s efforts? Those are finer grained. As x86-based servers established positions in key markets, Microsoft was a major beneficiary. But not everyone wanted the company to dominate data centers as it did PCs and by supporting Linux as a viable alternative, IBM and others were able to support healthy market diversity while sticking a thumb into Microsoft’s eye.
That passive resistance had a significant impact; IDC research puts the current market share of Windows Server at nearly 47%, a far cry from Microsoft’s 90%+ share of the market for desktops/laptops. But Linux’s rise likely triggered other Microsoft stumbles. With the exception of its own Azure platform, Windows Server lives in a tiny fraction of fast growing public cloud and/or hyperscale data centers. Linux is the dominant OS there, as well as in cutting edge supercomputing and high performance computing installations.
IBM gained significant commercial benefits from its Linux-related efforts, especially on the mainframe. Linux on IBM z Systems has enjoyed a 45% compound annual growth rate (CAGR) during the past decade, and now represents over a quarter (27%) of all mainframe capacity. Moreover, Linux is being used by four fifths (80%) of IBM’s top 100 mainframe clients. You don’t get more mainstream than that.
In other words, the bright future the company envisioned for Linux has truly come to pass but there is also room for improvement.
IBM’s Linux efforts away from the mainframe have been less stellar. The company’s sale of its System x group to Lenovo ended sales of Linux on Intel servers. More importantly, various IBM efforts to develop a sizable Linux business for Power Systems offerings faltered or fell away. However, the company revived that strategy last year with new POWER8-based systems with integrated Linux. Plus, the decision to open source its homegrown POWER processor architecture via the OpenPOWER Foundation was deeply inspired and informed by IBM’s years of successful open source collaboration.
LinuxONE – Linux without limits
So what are we to make of the company’s new LinuxONE systems and Linux without Limits strategy? In essence, IBM hopes to both extend and expand the benefits it has gained during a decade and a half of working with Linux. The new solutions are aimed at discrete audiences; the massively scalable Emperor at IBM’s enterprise mainframe customer base, and the modestly-sized Rockhopper at smaller organizations, along with new markets and use cases. In fact, these configurations mirror earlier mainframe iterations consisting of enterprise and “midmarket” or “business class” systems.
The new LinuxONE servers take advantage of robust current generation mainframe technologies but major differences lay in IBM’s software and licensing schemas, both of which are far more flexible than previous generation mainframes. In short, the new systems are entirely free of IBM’s z/OS, and can be configured as bare metal, bare metal with containers or fully virtualized environments with a range of OS, hypervisor, management, database and analytics technologies.
That remarkable flexibility extends to use case/financing arrangements, including pay as you go pricing, fixed monthly usage, variable pay per use and a new 36 month lease that can be cancelled after the first year. Some of these options have been available in past IBM z offerings but never all at once or with as many variations. Linux without Limits, indeed.
IBM is also renewing its dedication to the Linux community with the new Open Mainframe Project, a collaboration with the Linux Foundation that involves IBM contributing its largest ever amount (more than 250,000 lines) of mainframe code to the open community. As a result, the Open Mainframe Project bears some resemblance to IBM’s autonomic computing initiative (originally called eLiza) which was designed to incorporate key self-managing and self-healing mainframe features across the company’s other server platforms.
The difference here is that these IBM mainframe features and functions will be available to the entire open community, a point that underscores both the value of IBM’s new Linux efforts and the depth of its commitment to open platforms. It wouldn’t be too far-fetched to call the Open Mainframe Project a renewal of open source vows. Relationships may change over time so it can be valuable for individuals to remember common values and reinvigorate their partnerships.
IBM has certainly worked hard to ensure the commercial success of its new LinuxONE offerings but by refreshing its commitment to the community-driven values at the heart of Linux it is hoping to, once again, catch lightning in an open source bottle. Those values have paid off hugely for IBM, and its mainframe customers and partners since 2000. If LinuxONE is endowed with similar levels of energy and dedication, IBM’s Linux without Limits could move from being an intriguing concept to becoming a common reality.
© 2015 Pund-IT, Inc. All rights reserved.