Application Center 2000 reins in administrative costs
IT managers running critical Web applications on cobbled-together clusters of Windows-based servers finally have a tool, Application Center 2000 Version 1.0, that takes most of the pain out of building and maintaining these clusters.
During eWeek Labs tests, we were able to accomplish all the common tasks needed to bring a number of relatively inexpensive machines into a cluster running Windows 2000 Server. These tasks included setting up load balancing, application management, content replication and performance monitoring.
Using MMC (Microsoft Management Console), we were able, for the first time, to manage and monitor clustered servers as a group.
IT managers looking for a reliable tool to manage critical Web-based servers and applications should check out Application Center 2000, which started shipping last month for $2,999 per processor. In these troubled economic times, Application Center 2000 should significantly reduce cluster management costs and boost application reliability. The tool now makes server clustering available for Windows 2000 Server, not just Advanced Server.
The major trade-off is that the product works only in Microsoft environments. Integration with third-party, load balancing products from vendors such as F5 Networks Inc. is still quite primitive and limited to grow-your-own API calls.
The other notable weakness in Application Center 2000 is the sluggish response from the MMC console. Although IT personnel probably wont be fazed by the poky response, we often found ourselves staring wistfully at the screen as we waited for the status of application deployments to be reported.
It was easy for us to dismiss these weaknesses, however, because of the nice action provided by the collection of wizards that are the backbone of Application Center 2000s cluster-building and management abilities.
We installed Application Center 2000 on a machine running Windows 2000 Server, and within minutes we were able to start building a cluster. The beta version we tested in September required either Windows 2000 Advanced or Data Center Server.
Other wizards made it simple to add or remove machines from the cluster and deploy applications. We were surprised at how few steps were needed to accomplish the once-herculean task of building a Windows-based cluster. Even new Windows NT administrators should have little trouble using the Application Center 2000 wizards to construct large Web server farms.
We found it generally easy to follow the Application Center 2000 documentationto start performance monitors and to keep tabs on events, for exampleeven where wizards were not supplied. Although most of the performance information we tracked was mundane processor, disk and memory utilization, it was handy to be able to track this information from a single console.
Other products and even free utilities in Windows 2000 Server provide this performance information for single machines, but we found it especially valuable because it was combined with other IIS (Internet Information Server) information and was easily seen across all servers in the cluster. For example, we were easily able to track CPU and memory usage, system and application events as well as TCP/IP port utilization and health monitor alerts. We were able to set threshold limits and were notified via e-mail and event log entries when they were exceeded.
An especially convenient feature of Application Center 2000 is that the product synchronizes server management policies across all selected servers in the cluster, thus ensuring that our policies were enforced without a lot of additional administrative overhead.
During tests, we quickly distributed network and component processing loads among the clustered servers. Application Center 2000 uses a weighted fair-queuing mechanism to divvy up the network load. This same process allowed us to remove servers from the cluster without disrupting user access to the cluster as a whole.
Application Center 2000 has a number of capabilities that are standard for server management products, including the ability to restart machines, execute script files if administrator- defined conditions arise and apply remote administration. These features worked well in tests, although they dont stand out from other management tools.
Cameron Sturdevant is the executive editor of Enterprise Networking Planet. Prior to ENP, Cameron was technical analyst at PCWeek Labs, starting in 1997. Cameron finished up as the eWEEK Labs Technical Director in 2012. Before his extensive labs tenure Cameron paid his IT dues working in technical support and sales engineering at a software publishing firm . Cameron also spent two years with a database development firm, integrating applications with mainframe legacy programs. Cameron's areas of expertise include virtual and physical IT infrastructure, cloud computing, enterprise networking and mobility. In addition to reviews, Cameron has covered monolithic enterprise management systems throughout their lifecycles, providing the eWEEK reader with all-important history and context. Cameron takes special care in cultivating his IT manager contacts, to ensure that his analysis is grounded in real-world concern. Follow Cameron on Twitter at csturdevant, or reach him by email at email@example.com.