Major Face Lift

By Cameron Sturdevant  |  Posted 2007-04-25 Print this article Print

SCCM 2007 requires Microsofts Active Directory, SQL Server (and version) and extensions to the AD schema.

We had some initial trouble installing the management point correctly because we tested SCCM 2007 Beta 2 on a system running Windows Server 2003 with SP2. Only SP1 with a few new patches passes the prerequisite checker. After starting our installation over with the penultimate version of Windows Server 2003, we were able to get the management point working.

Although SCCM 2007 code was substantially rewritten to include many capabilities that were add-ons in SMS, it is possible to upgrade from SMS 2003 to SCCM 2007 to preserve collection groups and much of the inventory information that organizations may have already captured in SMS.

Microsofts acquisition in 2006 of inventory specialist AssetMetrix has substantially improved software identification over previous versions of SMS. IT managers who upgrade to SCCM 2007 should re-run most software and hardware inventory jobs so that SCCM 2007 can take advantage of the plethora of new data for determining which machines are Vista-capable and what software applications are in use throughout the organization.

SCCM 2007 uses its revamped inventory collection capability to enable greater license compliance. For example, during our tests, SCCM 2007 was able to distinguish between OEM and MSDN (Microsoft Developer Network) versions of Office.

Much of what was error-prone and tedious in SMS 2003 is now automated or wizard-driven, especially the task sequences in the now-integrated OSD (Operating System Deployment) module. With extensive advice from Microsoft support, we used the Task Sequencer to install Vista, configure network settings, migrate user state information and install updates and application software.

Once mastered, Task Sequencing will save IT staff significant time when it comes to deploying systems. However, expect considerable upfront learning and tuning time to get the most out of the feature.

SCCM 2007 preserves much of the look and architecture of SMS, including primary and secondary sites. However, a big change in this iteration of Microsofts management tool is the addition of six new server roles that improve the scalability of the product. All these server roles—which include management, server locator, reporting and distribution points—can run as services on a single Windows server. However, for performance and scalability, many of the services, including those used for distributing applications and OS and collecting inventory information, should be installed on separate servers.

To read more about how eWEEK Labs tested SCCM 2007, click here. We ran some of the new server roles on machines in the DMZ in our test network. For example, System Health Validation is a new Internet-facing function that checks the status of endpoint devices, including installed software and firewall settings.

Other new server roles, such as the fallback status point that we used in user state migration tests, can be distributed throughout the organization to improve new system deployment efficiency. Even though the services can be spread across organizations, they are managed from a primary site server for centralized reporting.

In our tests, the primary site server stored data for itself and all the sites beneath it in a SQL Server database. A primary site server has administrative tools that enable direct administration of the site.

In particular, the ability to perform bare-metal installations, such as those likely to occur in a centralized depot setting, is much improved in SCCM 2007. Using WDS (Windows Deployment Services) and a PXE (Preboot Execution Environment) server, we were able to advertise OS deployments for almost hands-free provisioning. Not surprisingly, because SCCM 2007 is in beta, we werent able to achieve the fully hands-off deployment that is promised in the final product.

But, even when the final version ships, SCCM 2007 will require the full-time attention of at least several IT staffers: Between building reference systems to create model images and testing deployments in either bare-metal deployments or in much more difficult side-by-side machine swaps, there are still many opportunities for errors to occur in an "automated" deployment.

Next Page: Improved security.

Cameron Sturdevant Cameron Sturdevant has been with the Labs since 1997, and before that paid his IT management dues at a software publishing firm working with several Fortune 100 companies. 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, with a focus on Android in the enterprise. 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 reviews and analysis are grounded in real-world concern. Cameron is a regular speaker at Ziff-Davis Enterprise online and face-to-face events. Follow Cameron on Twitter at csturdevant, or reach him by email at

Submit a Comment

Loading Comments...
Manage your Newsletters: Login   Register My Newsletters

Rocket Fuel