Back in 1998, several members of the Silicon Valley Linux Users Group (
From mid-2000 through early 2002, most news about LinuxCare run under headlines like these:
Lessons from the Layoffs at Linuxcare (2000)
Surviving the downside, Linuxcare remains focused (2001)
Linuxcare finally gets a permanent CEO; is he in time to save the ship? (2001)
Fundamental business decision paying off
That “permanent CEO,” Avery Lyford, is still there, and so is LinuxCare, so apparently he was in time. And the company that went from a high of about 280 employees to fewer than 50 is hiring, albeit cautiously. And, says Lyford, the decision to get out of the Linux support business and start selling a software product instead is what made the difference. “Support has become a good market for people like IBM and EDS that have big support businesses, and even for Red Hat,” he explains, but he doesn’t think it’s right for companies the size of LinuxCare. Besides, he admits, “I’m more of a product person.”
Lyford has described himself as “…a geek and proud of it” in earlier interviews, and also has a well-developed sense of irony, as his description of his start at LinuxCare makes clear: “September 10, 2001 was a great first day. The second day was a bit more difficult, being 9/11.”
Obviously, since that bleak point in U.S. history, things have picked up for both Lyford and LinuxCare.
The proximate reason for the change in corporate direction, from service to software product, says Lyford, is that the first thing he did when he joined the company was “interview a bunch of customers, and ask, ‘What are the obstacles to adopting Linux in the enterprise?'”
This led to the idea of emphasizing Linux in corporate data centers and ruminations about how the company could leverage its friendly relationship with IBM.
Talking to customers proved valuable
With these thoughts in mind, Lyford says, “We interviewed data center customers and found wide-spread interest in Linux, but what came back overwhelmingly was a need for tools, not support — the ability to roll back (software changes and updates) and increased provisioning efficiency to make it easy for people to use Linux.”
So Lyford decided LinuxCare’s best bet would be to provide tools specifically for managing Linux in data centers, where the biggest expense is neither software licenses nor hardware, but people, who typically absorb around 60% of corporate data center budgets.
Develop a tool to make those people more efficient, and you would have a hot-selling item, right? But the need went beyond making life easier for sysadmins trying to install or update software on many machines or on many partitions on a mainframe. Lyford soon realized that another big sales plus for whatever tool LinuxCare developed would be the ability to make use of people corporate customers already had available, including Windows-trained administrators with only point-and-click experience. In other words, while anyone running a data center of significant size would certainly need to have some seriously skilled engineers around no matter what, couldn’t a lot of routine, repetitive tasks be scripted so that they could be performed through a GUI by lower-skilled, less-expensive workers?
Then came discussions about culture differences. Linux, Lyford says, “tends to be about adaptability and and flexibility.” At the same time, he points out, “Data centers tend to be about stability and reliability … the trick is to get an infrastructure both flexible and reliable.”
You’re always dealing with a heterogenous computing environment
Not only that, even if you say “Linux” all day long, and your company is named “LinuxCare” and you are personally a Linux booster (as Lyford is), you must deal with the fact that major corporate data centers are a heterogenous environment, with Unix, Windows, and older mainframe operating systems all over the place.
Lyford puts it this way: “Here’s the dynamic that drives CIOs crazy: How many applications you have to support, how many platforms you have to support, and the big one is, you know, your budget.
“It’s frustrating. People (in the company) will come up with a new application they need, but it only works on a certain platform.”
Ah! What about those big mainframes that can run many partitions with a different operating system image in each one? What if you came up with a Linux-based utility that made it really, really easy to bang in new partitions whenever you needed them? You could have Windows, Linux, and Unix all running side-by-side on the same zSeries unit, and if your utility was simple enough to use that even a newly-minted MCSE could create and manage those partitions (with, hopefully, high-end engineers in the background to step in and handle the hard problems), you’d probably have a pretty salable software product, right?
What if you made said product so simple to use that the putative MCSE newbie could create and provision a new partition (essentially a virtual computer unto itself within the mainframe) in only a minute or two?
Sure, you need to have a high-level person around to create the templates that allow this to happen, but you only need to create each template once, and from there on it’s point and click all the way. Hey! Make it that simple, and maybe you don’t even need data center people to set up virtual servers. You can simply assign each internal client department a certain amount of disk space and tell them, “Go to it. Set up all the virtual servers you want, with any operating system and applications template we have around, and give us a call only if you have a problem or need a new template created.”
Centralizing all those servers
In CIO-land, this is nearly nirvana: Hundreds or thousands of user-created, user-administered servers, all running on one big mainframe that takes a lot less maintenance time per server than having all those servers housed in hundreds or thousands of discrete boxes.
Not only, that, Lyford says, users like this system too, because, “there are a whole lot of people in the world who don’t want to learn a lot about servers, but just want to go about their business. Applications developers, for instance. Today they have to get someone else to provision a server for them. With Levanta, they can serve themselves.”
I don’t know about you, but I’m sold. All I need now is a zSeries mainframe to replace my little Toshiba laptop …
Apparently many corporate data center managers and CIOs who do have mainframes to worry about find LinuxCare’s Levanta attractive, and enough of them are buying or at least thinking about buying the software that LinuxCare is growing instead of shrinking — they say they’ve hired 15 people in the last year — and is doing well enough financially to satisfy its board of directors and the investors behind them.
No one is trying to hide the company’s up-and-down history. LinuxCare’s press center page carries a fairly complete archive of press releases and news stories (including some negative ones) dating back to the company’s beginnings. Their resource page is about as Linux-oriented as you can get, and so is their Community Center page.
The open-projects.linuxcare.com is sparse and out of date (and is not currently linked from the LinuxCare main page), but Lyford says LinuxCare employees actively contribute to many Open Source projects, and we can expect to see some corporate-backed Open Source releases “soon.”
Company focused on a single product
For the moment, LinuxCare is focused on its single product, but there may be extensions of it for non-mainframe use available before long.
There are no plans for an IPO in the immediate future, nor are there unrealistic performance expectations from investors, says Lyford.
What started as a razzle-dazzle, ’90s-style “Linux hype” company seems to have settled down happily under mature leadership, and has turned into a focused, albeit proprietary, software company that shows strong potential for slow but steady growth in years to come.
