Open Source Meets Systems Management
Traditional systems management software from the Big Four enterprise-class systems management players BMC, CA, IBM and Hewlett-Packard sought to solve a difficult problem: the monitoring and management of the diverse hardware and software systems that make up corporate IT. The promise was that these systems-management suites could keep tabs on and manage systems regardless of the vendor, underlying chip and operating system. As a result, they were complex, pricey and hard-pressed to keep up with changing requirements. While Microsoft offers Windows-centric systems management and VMware offers management tools for the virtualized world, the Big Four are really the incumbents to watch as more open-source systems management solutions come online.
Here, Bernd Harzog, analyst with The Virtualization Practice LLC, talks about how newer, more agile open-source alternatives are changing that.
Q: More companies are using open-source tools to monitor and manage IT. What are the names to watch?
Harzog: There are some great products with open-source elements. Hyperic, for example, is used to monitor very large-scale Web environments. It was acquired by SpringSource, which was then acquired by VMware. Hyperic has a yet-to-be-publicly-determined role in VMware’s monitoring strategy. It competed with Nimsoft before Nimsoft was bought by CA, and with Nagios, which remains almost a pure open-source -- or at least open-at-the-edge -- product.
Q: What’s the appeal?
Harzog: If a systems-management product is managing a customer’s current data center hardware, it also needs to manage new equipment the customer adds over time. Once you have more than five customers, it’s impossible to keep up with that. The only solution is to have systems-management software that is open at the edge and that is extensible and adaptable, so anyone can build a “collector” that collects the data about the new target device. Companies like CA, HP, IBM and BMC are on 12- to 18-month release cycles, so when something gets popular, it goes on the roadmap, but support for it will take up to a year and half.
An open-source-oriented company like Zenoss is open at the edge. Anyone can build a Zenoss ZenPack collector or a Nagios collector for new devices and not have to wait for the vendor.
Q: What else drives demand for these systems-management tools?
Harzog: Virtualization, IT-as-a-service, public and private clouds. Those things tend to break legacy products. Virtualization introduces new requirements -- namely keeping up with dynamic systems. Legacy tools are not designed to do this. They can monitor virtual machines (VMs), but they don’t do a good job keeping up with change in dynamic virtualization environments. Their approach to systems management, which worked with physical servers, doesn’t work well with virtual servers. The desire of enterprises for something more flexible, less expensive, easier to manage and able to meet the need of new-use cases drives demand for these new tools.
Q: How about systems management versus systems monitoring?
Harzog: That’s the other side of systems management. Management includes updating, configuring, provisioning things. There I would look at interesting new tools, like Puppet Labs’ Puppet, Opscode’s Chef and ScaleXtreme.
Management has to be done correctly. It starts with provisioning and configuration management and then has to manage performance and availability all up and down the entire stack. It’s a hard job, and tools are evolving quickly.