I first met Javier Soltero (CEO) and Doug MacEachern (CTO) from Hyperic last February. They were at the point of determining how best to go large with what was otherwise a cool concept, and as ex-Covalent, Apache-shaped people open source software was obviously to be part of the mix. Hyperic is one of the companies that Tim O'Reilly identified in an Executive Briefing session at OSCON 2006 as a company to watch.
Hyperic started as an off-shoot from Covalent a couple of years ago. Covalent was reshaping itself, and this intrepid team gained the rights to the technology base they were building around large scale system administration and started their own company. They've been profitable almost since they began with key customers and an early OEM relationship with JBoss. (There are further impressive customers than those listed that can't be named.) They're now funded and away to the races.
The list of cool things to love about the company and the business model:
- It recognizes that the HQ development team can't innovate all the plug-ins to keep up with all the customers scenarios and is creating and investing in their user community. This value provides value to all the other users and customers in the HQ world. Their Free Software model enables this.
- Sys admins also dig deep into how things work. They're responsible for maintaining the integrity and stability of their facilities and have learned to not trust vendor claims of omniscience and omnipotence. They're the first people answering the phone when it all breaks. Providing the software as Free Software supports their customers needs (and fears) as they in turn support their end customers.
- Those same knowledgeable sys admins can help harden the underlying tech base in their maze of twisty environments (all different) as it's Free Software. We all know that the ratios of users to bug reports to bug fixes to well-formed patches falls by orders of magnitude when you "run the numbers" over history. But history doesn't measure when the diamonds arrive, and community led innovation isn't being disdained or marginalized by Hyperic.
- Hyperic recognizes that their potential customer base is a well connected group of people that share software and ideas constantly through multiple channels, and gives them something to share. System administrators have always believed in open source software (for decades before we called it open source) and have always shared tools and knowledge. Hyperic understands their customer.
- The technology base started life recently as a well thought out re-engineering of the problem space, built over a couple of years by a coherent well-funded team. First, contrast this with the Frankenstein products of the Big Guys in this space (BMC, CA Unicenter, Tivoli, HP), stitched together through acquisitions to try to solve more of the problem. To now take the product into community means there's a stable core around which the community can rapidly contribute. (I am assuming the community engagement will be well managed here as John Mark Walker is in charge.) This is a much shorter life cycle to critical mass from a business perspective than starting as an open source software project that needs to build the core base over time before trying to build a company around it.
- Attaching the GPL to the open source software base means the Big Guys won't be able to take advantage of the published source base, while customers and the user community can. N.B. I did not call these players competitors. HQ might be competition for their offerings, but they are not competition for HQ. Hopefully Hyperic stays laser focused on their customers and user community, and ignores the distractions from the Big Guys, just like MySQL did.
Prediction: the Big Four will now start the rhetoric around open source with respect to security, scalability, and reliability that we saw first from the OS vendors, and then the database vendors.
- The base open source is completely useful by a sys admin. The enterprise subscription provides support, training, indemnifications, certified binary images, and a few additional features that a really large IT environment would want to add.
- Hyperic has a flat pricing model and one that is inexpensive enough to creep in under the floor boards of large organizations. Contrast this with the more costly Draconian multi-axis pricing models of some of the Big Guys which essentially penalizes the customer for each new management point when they need it the most. This literally becomes self-limiting. There will come a day when the customer says "enough". (And the day will probably be extended with new enterprise agreement style subscription pricing to try to assuage the customer and extend the life of the Big Products just a little longer.)
- Instead of engaging in a heavy expensive sales cycle with the C-level execs against the Big Four Incumbents around Enterprise Management, they can slip into the bottom of the organization buried in departmental budgets when the open source users become customers. They can send alerts up through the big tools, so they fit into the environment quietly and can begin to displace it.
I finally got a demo last week at Linuxworld. I am not a system administrator or network ops person, but I was standing with two that run really big networks and both saw the potential. (One works in a Really Big Facility with an installed base of tools that would need to be gently introduced to HQ, but he was still impressed.)
If you're a sys admin or merely count them among your friends, take a look at Hyperic HQ here. If you're a business person thinking about open source, start imagining how to engage a community of users in the small (away from the internal toxicity of ROI and TCO discussions with the CxO and an incumbent sales force), so they grow into customers in the large.
Congratulations, gentlemen. Well done.
Hi Stephen,
I _am_ a sys admin and I took a look at Hyperic HQ a few months ago -- I liked it, but the enterprise pricing is too steep - $780 a year for each machine, up to two CPUs. Add another $780/yr. for each pair of processors in a given machine - $1560/yr. for, say, an IBM x365 - that's a lot of money for a mid-level server, never mind the big SPARC DB machines. (An enterprise license is needed for just about any environment that would have to deal with SOX or HIPPA, for things like RBAC.)
A few hundred servers (or fewer if they're 4- or 8- way boxes) and you're at a quarter-million bucks, annually. At that sort of money, suddenly the C-level execs start asking "why not (Tivoli/OpenView/whatever)?". If they're spending that much, most will be wary of some "new guy".
My former employer didn't even consider the product due to the price. Something like Manage Engine (http://manageengine.adventnet.com/products/applications_manager/index.html) was coming out far cheaper, with similar out-of-the box functionality. I liked Hyperic's design and think it would be more flexible in the long-run, but a Tivoli replacement it is not (money-wise, that is).
Posted by: Ben | 21 August 2006 at 07:48
Hi Ben,
Thanks for the feedback. Our website notes that larger environments such as yours may opt to obtain a site wide, enterprise subscription which is not done on a per machine basis. We'll make sure to make that clearer on our documentation.
-javier
Posted by: Javier Soltero | 21 August 2006 at 09:50
Hi, Stephe,
Yup, your blog is on my list of RSS feeds. Call it nostalgia for the days when I had to worry about this sort of thing...
My eye was caught by your phrase: "away from the internal toxicity of ROI and TCO discussions with the CxO and an incumbent sales force". I don't think you're being entirely fair here. ROI and TCO measures are, it seems to me, entirely valid ways of comparisons between alternatives. Indeed, if the new tools can slip into the existing infrastructure as symbiotes, and then demonstrate that they are reducing TCO, then I'd be all for them. I would agree with you if what you meant was that the sales force of the established vendors had internal proxies in the customers company. That can certainly be toxic. But ROI and TCO? If, in your view, these measures themselves are toxic, then what you you put in their place?
Posted by: Geoff Coupe | 21 August 2006 at 10:24
Good Lord -- It's Geoff Coupe.
You are correct in your interpretation. I've seen too many internal proxies use (for example) things like the Microsoft TCO arguments around Linux. Those TCO studies were missing critical cost items but if the exec hasn't the time to detail dive, then the summary numbers can get the vote.
I'm worried (and your opinions and experience are relevant here) that the vendors have focused the discussions too much of late on the cost side (features and pricing) and indeed scoped ROI that way. The business value (versus IT cost or cost savings) seems to have become lost in the noise since fiscal restraint and lowered budgets arrived. The business value in how a system serves the company's customers would seem to be more valuable to the bottom line. (I might need a separate post here, because that's a reasonably large discussion on its own.)
Posted by: stephe | 21 August 2006 at 10:44
Stephe, I agree with your view that business value is the way to measure, and TCO should include such elements. It is, of course, often difficult to measure objectively, but that's half the fun :-)
Posted by: Geoff Coupe | 21 August 2006 at 22:10