Never mind fat-bellied tech titans, give enterprise upstarts a chance
Why IT shoppers miss real bargains
Sysadmin blog In the IT world, momentum is everything. The past few months of talking to various start-ups have been an eye-opener for me; but none so much as talking to Bill Karpovich of Zenoss. Zenoss's story reflects one I've heard from many other start-ups of late; they have great software and are growing rapidly, but there's always that risk that being good at what you do might not be enough.
Zenoss has open-core monitoring software; the open source version has a reasonable following with the enterprise interface offering some nice features that fit well with modern "treat IT as a service" trends. They're big on root-cause analysis and displaying monitoring as applications and services rather than systems. The canonical example is a website that is malfunctioning, but still "up". In this example, the web servers, underlying hosts and network switches are up, but the system is non-operational because the SQL server bought it. Zenoss's kit analyses the failures then offers up a list of possible sources for the error, with a likelihood of that specific source being the cause attached.
Cool idea and I rather like their implementation. While I am personally convinced that Zenoss is ready to shake up the enterprise monitoring market, my completely unscientific poll of 12 1000+ seat deployments sysadmins saw none of them willing to give smaller vendors a look. The question for the world's alternative vendors is: how to get enterprises to bite?
In a couple of cases, I've managed to needle the admins in question long enough that they caved and gave took a look at some vendors I think could work to their advantage. While I admit to some evangelising about neat tech on my part, my prime motivation is usually a desire to test-drive new stuff on test labs and production deployments that are an order of magnitude larger than I can bring to bear on my own.
When I can get these sysadmins to look past their preferred vendors, they discover that start-up entrepreneurs have learned a few tricks since the turn of the millennium. (Which is when a lot of their prejudices were cemented.) Start-ups are producing enterprise-class software; it is increasingly rare for the more popular start-ups to be making stuff that works fine for SMEs, but falls apart at scale. Despite this, enterprises remain conservative; they want a lot of someones to walk through perceived minefields in front of them, sometimes writing off entire segments of the market based on bad previous experiences.
The traditional approach is simple: target SMEs. Unfortunately, a monitoring company like Zenoss - regardless of whether or nor you're a startup or a tier 1 - is that the raw diversity of devices and applications pose a barrier to entry. Even were Zenoss inclined to write support for each application and hardware widget in the SME world, the costs would be astronomical. To build out the support needed for the sheer breadth of applications and devices in the SME world, community-provided plugins are essential.
Some enterprises will inevitably look at the interdependence of Zenoss with an open source community and worry. The relationship could blossom - Apache and Red Hat are great examples - or it could sour. (Just look at anything community based that Oracle has ever touched.) The community could lack the skills, vendor contacts or interest in supporting some critical element of enterprise infrastructure.
There's some irony to be had there; the enterprise side of this is comparatively easy. Big vendors publish APIs and even work with you to make sure your stuff interoperates. Monitoring companies - and start-ups in general - will have an easier time today working in an enterprise environment than elsewhere; yet the conservative nature of the enterprise can remove them from consideration.
In the specific case of Zenoss, I don't think there's much reason to worry. I haven't yet run into anyone who has used the enterprise product that has complaints. The open source community surrounding the core product is enthusiastic and upbeat; Zenoss' community has proven to be an asset, not a liability. The negativity on my part is a reflection of my own deep cynicism. I look at the technologies on the table - from Zenoss and other start-ups - and think simultaneously "this is better than the tier 1 stuff" and "it's going to be like pulling teeth to even get a foot in the door". For some, the route to acceptance is to be bought out. For an increasing number of open core outfits with a GPL licence, that's not a very likely option. Many start-ups will survive or fail on their own. For Karpovich's team, it's Zenoss against the world.
It is perhaps heresy to discuss such things on a technology website. We like to believe in the myth that good technology conquers all, but a myth is all that is. Just this once, I'm to cheer on the marketing and sales bodies - I selfishly want them to succeed. I lack the 7+ digit budgets required to do things the tier 1 way. I want Zenoss to succeed; they make my life easier and allow me to get more done for a cheaper price. Ultimately, isn't that the point of IT? ®
Sponsored: Benefits from the lessons learned in HPC