In this interview with SearchEnterpriseLinux.com, Karpovich lays down the case for open source alternatives to IBM Tivoli and Hewlett-Packard OpenView, and why "light" versions of those huge -- and expensive -- applications suites have a bright future in the mid market space.
How does your product compare with those from IBM, HP and others and how does it benefit the IT manager?
Bill Karpovich: We're developing an open source systems management platform that is easier to use. It's a platform that is open, and compared to other open source applications, more complete. We wrote Zenoss from the ground up much in the same way as those proprietary products from IBM and HP began. We wanted to develop an enterprise-class application.
Along the way, we had to leverage some of libraries and components already out there, but at its base Zenoss is something that we raised to be used in the enterprise. With vendors like IBM, IT managers are receiving very large, very expensive products, but what they do and what people use Zenoss for are very common tasks like monitoring. Monitoring is nothing new, but because it is open source our customers are free to customize it as much as they'd like.
If a customer forks the code, or adds a custom plug-in to their stack, is there a danger that they can no longer be supported? Do you make this clear when working with them?
Karpovich: This is something we need to be thoughtful of when speaking with customers. Zenoss software monitors networks, servers, applications and services. The biggest benefit, however, is its openness, meaning that users can tailor it to their systems any way they choose. With these products, most of the customization done by IT managers is done on the edges with well defined APIs. They are not modifying the source code; they are modifying the user interfaces. They are building plug-ins, and using Web services APIs.
You've spoken to our sister site SearchNetworking.com in the past, and said you compete with IBM using an 80/20 approach. Explain a bit more what you meant by that.
Karpovich: When you're talking about just monitoring that is correct. An easy way to convey what that means is to say Tivoli and OpenView have a lot of functionality, and we don't go into some of their fringe features with Zenoss Core. However, the nice thing about open source is that when IT managers have a need for those features in the future, they will evolve and change the software on their own platforms.
So what's in the 20% fringe Tivoli and others offer that IT managers don't need right now?
Karpovich: With these big platforms, they have focused resources on compete monitoring applications for less pervasive technology. This could include monitoring a specific application like SAP, for example, whereas we don't have specific SAP monitoring in place. This would be more appealing to big companies that are going down a list with a very comprehensive set of boxes to check. These are customers with very large enterprise infrastructures. However, once you get into the mid-market companies with thousands of servers the need is not as broad. For example, in Microsoft Word I've paid for whole package, but I've never used Mail merge.
You've claimed in the past open source monitoring is gaining traction -- mainly with high download numbers -- but can you give more concrete examples? Customer wins for example?
Karpovich: There are a couple of ways to talk about traction. We've mentioned the downloads and SourceForge project of the month for March recognition [on our website]. What's really been exciting is we increased our rank from 154,000 to the top 20 over the past year, and we're on track for 40,000 downloads this month. The project is well into the business-to-business stage; we're not talking about downloading the latest IM client.
Overall, this speaks to the penetration in the market for open source applications. Relative to Nagios, which is a first generation open source product, they've done a nice job, but there is no company behind it. The maintainers never thought of it as a project that an IT manager would use to monitor an entire enterprise environment. What we've done is bring that type of product to mainstream adopters. In the past few months, we've actually acquired 10 new commercial clients including several billion dollar companies.
People compare Zenoss with similar players in this space, mainly Hyperic and GroundWork. How are you different?
Karpovich: Companies like GroundWork are similar to the Red Hat approach, where a company gathers up the pieces and puts support behind it. Our approach is we have always had the code and we are in control of its roadmap and indemnification. The Hyperic model is where a company comes from a commercial background and makes some of the code open source.
Some IT managers are apprehensive about open source support problems. How is that issue being addressed with systems management?
Karpovich: There can be perceptions that open source software is hurt by poor support, but we think of ourselves as a commercial open source company. We compare ourselves to the Linux model that took off when a popular company like Red Hat got behind it.
There have always been open source management alternatives available, but there hasn't been a company that got behind it to provide support and legal protection. And even if a project does go under, which people today see as the worse case scenario, the product code will always be available because it is open source. A community will still exist to provide worldwide support whether the commercial company is there or not.
Have a question or comment about the article? Email Jack Loftus, News Writer.
Be sure to check out our new blog at the Enterprise Linux Log.