Advice for Open Source Startups: Remember LinuxCare 116
LinuxCare
was born in 1999 -- venture-backed by top tier VC firms like Kleiner
Perkins, with total
funding in the ballpark of $70 million.
Those were the frontier days for Linux. There was
a ton of industry interest and activity despite the fact that the jury
was
still out with respect to end user adoption. Nobody really
knew exactly
how Linux was going to be used – would it be for the desktop,
servers,
etc.? The company used the vast venture coffers to promote
the brand and
staff star-power (even Linus Torvalds consulted for them
briefly)– and
LinuxCare quickly became the recognized name for Linux services and
support,
doing work for big systems vendors like Dell and IBM in addition to
developing
device drivers and offering education services.
Red Hat had the Linux OS and
software, VA Linux had the
hardware – and LinuxCare had the services. It was a
theoretically perfect
enterprise Linux ecosystem triumvirate.
But it wasn't meant to be.
The demise of LinuxCare can be attributed to many
factors. The first was that enterprises were slow to adopt
Linux
– in the early ‘00s, IT spending came to a grinding
halt
with the dot-com and stock market crash. But the key factor
to
LinuxCare’s spectacular death spiral was the fact that they
were going up
against Red Hat, the very company they were basing their business on.
Red Hat
not only developed their own distribution of Linux, but also started
offering
support for it. Red Hat offered a one-stop shop for Linux software and
services
regardless of hardware. Enterprise
customers decided it was easier to buy from one vendor. This same
sentiment is
what drives sales of Microsoft software in enterprises today.
LinuxCare suffered a painful public death over months
of executive departures and layoffs, VA Linux abandoned hardware for
software,
and RedHat, with the cash to weather the tech spending downturn,
expanded its
revenue streams and became the de-facto enterprise Linux distribution.
It's easy to dismiss LinuxCare as "ahead of their
time", which is definitely true. But the fundamental and fatal flaw was
that they based their products on someone else's IP, with no IP of
their own.
When the market tanked abruptly, LinuxCare didn't have the money to
weather the
storm and didn't have consistent alternative revenue streams to combat
the lack
of services income.
Some of the executives from LinuxCare went on to start a new company called Levanta, which focuses on Linux systems management. They have since developed IP in software and hardware that can sustain the business beyond the services revenue.Their LinuxCare experience taught them how to build a sustainable technology business model on top of open source software. No longer do they rely on IP that walks out the door every night in their employees' heads.
In the end, it all comes down to
IP. Building a
business on top of something you don't own is extremely risky.
Companies need
to develop their own IP to be innovative and have competitive
differentiation.
And if they don't develop it themselves, they need to acquire or
license the
relevant code to protect themselves and ensure they aren't caught
without
alternatives.
An Open Source Danger
Zone?
In my eyes, the bubble associated with open source is less related to the millions of VC dollars and more related to the reliance on software and components that are not part of a company's internal IP. When Oracle acquired InnoDB, it had a less than positive effect on MySQL, but MySQL is a smart enough company to not bet the farm on something it doesn't own. It owns enough IP to sustain its products-and it's business from the risk associated with relying on someone else's code.
IT Groundwork has built a business
on top of an open
source network monitoring project called Nagios. They don't own the
copyrights
and they don't employ the creator. Kleiner-backed SpikeSource offers
"certified stacks" of open source software components, but they don't
actually create the open source components themselves.
And in SpikeSource's case, Red Hat announced that
they too would offer "certified
stacks." Who do think is going to win that battle? Red Hat, the
one-stop
shop that offers the OS and the apps, or the company that offers merely
a
portion of the total package. Does SpikeSource have the IP or
alternative
revenue sources to withstand Red Hat? Let's wish them
luckand hope they
know the LinuxCare tale.
If there is a bubble, it will burst when the open
source projects these new company's products and services depend on go
private,
fork, or get acquired. The market for open source is so new we haven't
seen
much of this yet. Only time will tell if the recently funded open
source
companies can build sustainable businesses, or if this grand experiment
will
result in a few 800 pound gorillas and many tiny monkeys.
Have something important to say to the Slashdot community? Email roblimo at slashdot period org with the complete article or an article proposal.
Re:open source business - oxymoron? (Score:2, Informative)
You can make profits from services and support, if not from the software directly. At least, thats the standard line you hear from most OSS companies. I don't think that the system is dominated by GNU.org, at least not in a negative way.
Many GPL'd products are doing just fine, MySQL as a major example.
The whole point of this piece was to show that investors-and many in the community are being naive in thinking that they can make money easily. It can be done, but it's not a guarantee. The companies highlighted are just a few who have the potential of being screwed because they don't own the code that they base their businesses on. Just like LinuxCare.
This is not to say it can't be done, but it seems oddly prevalent in open source.
No, nice try though... (Score:3, Informative)
Red Hat didn't have IP for the most part. They were one of the only real "expertise" plays in the game at the time- and they had what it took. LinuxCare didn't have as much as Red Hat and had internal management problems, so it died. VA Linux bought it because they premised everything on being a Linux hardware play and then when the big boys twigged onto doing Linux support for their bigger iron, VA hadn't gotten large margins on their stuff and had overspent themselves on other things- AND couldn't fill the flippin' orders they had in hand towards the end.
Re:The article is rose tinted regarding Linuxcare (Score:3, Informative)
Re:What exactly to they mean by "IP"? (Score:5, Informative)
> geek-founded company taking on Professional Management
> to keep the VCs happy and getting royally fucked over
> by the Professional Management with the Impressive Credentials.
Bingo. I joined Linuxcare in early 1999 as employee #27, and stayed until January 2001. For a while it was paradise. I mean, who wouldn't like having Rasmus Lerdorf and Andrew Tridgell working a few cubes away? The level of talent there was really rather spectacular.
But then . . . . a couple of very bad VC-installed apples at the very top of the company destroyed the place by a) squandering the funding on absurd levels of growth and infrastructure, b) failing to IPO in time, and (worst of all) c) creating a geeks-vs-suits culture that came to consume everybody in petty office politics. In a reversal of stereotype, it was the original founders who had the most sanity. They held the tech talent in place as long as they could.
Sure, Linxucare's open-source services model was a little ahead of its time but the company, leaned down, could have hung on through the bursting of the bubble and eventually come to thrive. Its confidence that Linux had a future in the enterprise has been more than vindicated. Now, it's IBM Global Services and similar scooping up all the Linux services income. Sigh, what could have been . . . .
Linuxcare had some really sharp guys (Score:3, Informative)
Andrew Tridgell (Samba, Rsync)
Paul Mackerraas (Linux on PPC)
Rusty Russel (iptables and lots of other kernel stuff)
Rasmus Lerdorf (PHP)
in Italy, we had Alessandro Rubini, Paolo Molaro (now doing some really good stuff with Mono), and a bunch of other talented guys. The group in Canada also had some very good hackers. In short, there were a lot of smart people there - I doubt I'll ever see anything like it again.
The problems were twofold:
1) The upper management. In particular: http://www.advogato.org/proj/DougNassaurWatch/ [advogato.org]
2) Not really coordinating all that talent. That was bound to be hard, especially given the times, when all the companies were fighting over bright people, but there wasn't really a focal point like Redhat had with their distribution.
Not that I buy the point of the article - Linux services are big business and are only going to get bigger. And...guess what? No one person owns Linux. Covalent does services for Apache Software Foundation software without owning it. It's open source, so it's not really a problem if your business model doesn't conflict (as in the Mysql case).
In any case, I got a good deal out of it - I came over to Italy to work with that group, where I still live with my Italian wife (as of this summer:-).