Best of open source in enterprise monitoring
- 11 September, 2007 09:04
Open source software has had a foothold in the enterprise monitoring sphere for almost as long as open source has existed. One only needs to look at the sheer ubiquity of small applications such as MRTG (Multi Router Traffic Grapher) and its RRDTool back end to see that. What we haven't had from open source is the big application -- the comprehensive, community supported open source enterprise management suite that provides the depth and breadth of functionality that businesses need and generally find in closed-source competitors. That is changing in leaps and bounds. In fact, open source enterprise monitoring solutions are evolving so quickly, we won't even try to declare a clear winner yet -- but we're working on it.
Not all that long ago, an entirely open source monitoring system would encompass software from five or six different open source projects. A typical configuration might include the very popular Nagios for monitoring, one of many available configuration front ends to make Nagios easier to manage, a plethora of Nagios plug-ins, a package such as Cacti for more advanced trending and graphing, and a few Cacti plug-ins such as WeatherMap and Thold. Experience indicates these packages together form an extremely powerful and infinitely malleable solution.
By the time the entire collection has been fully configured and integrated, however, someone has spent a significant amount of time getting it completely dialed in. Any addition of new resources to be monitored usually requires fairly intimate knowledge of how things were assembled to begin with. Worse still, upgrades can be complicated due to the interdependency of various components. None of this would seem particularly arduous to a reasonably experienced Linux/BSD user, but it can be downright horrifying for someone without Linux or Unix experience. That fact alone has prevented open source monitoring from penetrating the average, largely Microsoft-based medium business.
All of that is starting to change. Within the past several years, major new, commercially backed, open source monitoring packages have been arriving on the scene at a dizzying rate. Some of these packages provide the glue necessary to coherently unite disparate packages that hadn't previously been integrated. Others have been grown from the ground up, their backers aiming at providing a completely unified solution. Major players in this space include GroundWork, Hyperic, and Zenoss.
GroundWork retains the nearly limitless flexibility of its underlying Nagios-based monitoring engine while adding a much more functional AJAX-based Web portal and configuration interface. Given that it is based on Nagios, it retains access to the multitude of available Nagios plug-ins and extensions, making it very likely that anything an average business has on its network can be monitored.
Hyperic is a purpose-built application monitoring package that can monitor very complex systems from the Web server, through the application server, to the database engine, and all the way down through the hardware and out into the network. All of this is done with an eye toward end-to-end performance. Hyperic doesn't have the kind of hardware device support or brass-tacks infrastructure views that either GroundWork or Zenoss do, but it makes up for that by being able to track and interrelate resources that are required for delivering a given application infrastructure in a very comprehensive manner.
Zenoss is one of a very few packages that has attempted to completely re-imagine the makeup of an enterprise-wide monitoring package. It has been designed from scratch to be fully unified and internally seamless. At the same time, care has been taken to retain the same support base that Nagios enjoys by including built-in support for Nagios plug-ins while also offering powerful monitor generation abilities. Zenoss's Ajax Web interface is one of the easiest to navigate and manipulate, and installation and upgrades are easy even for a Linux or Unix neophyte.
Another project to keep an eye on is OpenQRM, backed by Qlusters. OpenQRM doesn't fall directly within the enterprise monitoring space, but is an open source provisioning solution that is squaring off against much larger closed-source rivals. OpenQRM provides a means to dynamically scale availableÂ server capacity based on load characteristics. It can monitor server load and deploy additional virtual servers into a virtual infrastructure as they are needed. It is compatible with VMware and Xen and fully supports automatic provisioning and monitoring of Linux servers, with some limited support for Windows VMs.
These four products are just a sampling of what's out there now and is a taste of what's to come. It is no longer a question whether open source monitoring and management solutions will displace the large, closed-source gorillas that dominate the market today, but how long the gorillas can hold on.
Matt Prigge is contributing editor to the InfoWorld Test Center, and the systems and network architect for the SymQuest Group
"バッグ ゴールドを持っていますハードウェアバッグ、シャネル スタンプされます常に金。Howerver, クイズ開発メソッドは相当この。決して 感じる誰か他人を示唆している通常。プラトンのクローゼットの中にある完全トップのすべての行非常に手頃な価格料金。"WikiLeaks Party closer to registering
Philip's 'smart' lightbulbs hit Australia
Philip's 'smart' lightbulbs hit Australia
Australia lags Mongolia in Internet speeds
Salesforce.com to buy Clipboard, shutting down service