Exchange Server Management and Monitoring Tools

Comparing six of the market's top products

[Editor's note: At press time, Quest Software has changed the name of its product from Foglight 3.0 for Exchange to 2MA.]

As a network administrator, your role is to ensure the smooth operation of your organization's Microsoft Exchange Server system. To confirm that the system is running, you simply ping the server every few minutes or use Performance Monitor to keep an eye on it. However, ensuring that the system is running smoothly is a bigger challenge. The Exchange Server management tools that Microsoft provides aren't helpful in determining why your message store is filling up at an exponential rate. And using these tools to create a historical trend analysis of your Exchange Server system is almost impossible.

Many products on the market claim to fill this Exchange Server management void. I tested six of the leading products: NetIQ's AppManager 3.4 for Microsoft Exchange Server, BindView's bv-Control 1.5 for Microsoft Exchange, Quest Software's Foglight 3.0 for Exchange, Hewlett-Packard's HP OpenView ManageX SMART Plug-In 4.0 for Microsoft Exchange Server, BMC Software's PATROL 3.4 for Microsoft Exchange Server, and Heroix's RoboMon 7.6 Exchange Intelligence Solution Set.

Some of the products I reviewed are tailored to Exchange Server; others offer optional configuration components that let you adapt the product to manage other applications. All the products I tested are competing for a sizable chunk of your budget.

At the functionality level, all six products are aggressively competitive. Each program offers the same core functionality: It lets you proactively monitor Exchange Server systems, and it provides analysis of the logged data.

In this review, I looked at the six products from an Exchange Server-centric perspective. The only features I tested were the products' Exchange Server monitoring and management capabilities. (Some of the products offer additional functionality, which helps explain the differences in pricing between products.)

My test bed consisted of several Windows 2000 Server machines and a handful of Win2K Professional and Windows NT Server 4.0 systems. I ran all tests on Exchange Server 5.5, and I tested features such as the products' ability to monitor, apply rule sets for server management, and collect and filter data into historical trend analysis graphs. Some vendors, such as HP and NetIQ, already offer versions of their software for Exchange 2000 Server. However, at the time of this review, most vendors' Exchange 2000-compatible applications were in beta testing.

AppManager 3.4 for Microsoft Exchange Server
AppManager has long been regarded as the premiere management tool for NT-based environments. Featuring tight integration with Windows, optional support for more than 30 server applications, and a flexible scripting language, AppManager can handle all your network-management needs. AppManager 3.4 incorporates beefed up support for backup management, more predefined policies than earlier versions, and integrated Active Directory (AD) support.

Installing AppManager's Exchange Server focused tool was fairly painless albeit slow. The program requires Microsoft SQL Server 7.0 Service Pack 1 (SP1) or later. After the setup program copied the software's files to my system, AppManager spent almost 20 minutes registering files with the system, creating a SQL Server repository, and starting, stopping, and restarting the necessary services. Of course, this process is only a one-time event, but you might want to think twice about installing AppManager during a workday.

The software employs a multi-tier client/server architecture, which makes it highly scalable. You'll probably spend most of your time working with the software's console. Each remote machine runs an agent that continuously reports back to the main server, and the heart of AppManager is its Knowledge Scripts, NetIQ's proprietary policy format. The company designed Knowledge Scripts with versatility in mind, and you can use them to passively monitor your applications or aggressively take action in the event of network problems.

AppManager includes an auto-discovery feature that NetIQ designed to probe only the machine on which you install the software. To monitor services and applications on client systems, you must install the AppManager agent on the machines. My network's core services are distributed among multiple servers; I dreaded the idea of manually installing AppManager on each machine. Fortunately, AppManager includes AgentInstall Knowledge Script, an elegant method of providing centralized agent deployment. This script works by installing itself as a client on the main AppManager server, then installing remote agents on each managed client. The script is automated, so all I had to do was provide valid usernames and Uniform Naming Convention (UNC) paths for the agents to use. A few minutes after I input this information, the script installed the appropriate agents on the remote systems. The agents then used TCP/IP to communicate with the AppManager server.

Unfortunately, NetIQ hardwired into the code the ports that AppManager uses, and I already had an application mapped to one of the ports that the software employs. Thus, I was forced to change the port number of my FTP daemon. Although this modification wasn't a major hassle, I would've appreciated the ability to define which ports AppManager uses to communicate with clients.

After I configured everything properly, I launched AppManager, which is a Microsoft Management Console (MMC) snap-in, from the MMC framework, and the software prompted me for a SQL Server login. I input the default systems administrator account and selected the proper repository. AppManager was then ready to go. Because NetIQ designed the software as an MMC snap-in, you don't have to spend much time trying to learn a new UI. As Figure 1 shows, AppManager looks like other MMC snap-ins but has a busier interface. For administrators who work on different platforms, AppManager includes a Web-based console, which lets you work with AppManager from any platform as long as you have a Web browser at your disposal.

NetIQ developers wrote AppManager's Knowledge Scripts in VBScript, which makes customization fairly simple. However, for those who aren't familiar with BASIC, AppManager includes many Knowledge Scripts.

To test AppManager's monitoring features, I created a large dummy file intended to fill up the disk space on my Exchange Server system. The software responded to the situation by sending me a warning. I made things more interesting by creating a bottleneck between a client and the Exchange Server system, which slowed the mail delivery time. AppManager had already been monitoring mail delivery, so it noticed that congestion was impeding the traffic between the Exchange Server system and the client, then notified me. Next, I killed the Exchange Server service to see how the monitoring tool would react. Not only did AppManager notice the stalled service but it reinitialized the service. AppManager includes a ServicesDown Knowledge Script that you can use to automatically restart halted services on remote systems.

AppManager's ability to generate reports doesn't disappoint. The Report Manager tool includes 18 preconfigured report types that range from connectivity data to server load statistics to mailbox information. The software generates reports from the data that the agents collect and store in the SQL Server repository.

To test the reporting features, I created a traffic report by selecting the Exchange Server Traffic option. Report Manager then browsed the SQL Server repository and created a line graph detailing the system's usage level. I selected the TopNMailboxes option to see how much space each user was consuming on the server. Again, the Report Manager dissected the repository and output a list of every user, the size of each user's mailbox, the number of messages in each mailbox, and the last logon and logoff times.

After you generate reports, you can export them to almost every file format available (e.g., Microsoft Excel, Microsoft Word, Lotus 1-2-3, HTML). However, the software lacks an option to automatically export reports as HTML documents and publish the files to Microsoft IIS; this feature would have been handy. To view reports remotely, I had to manually copy the HTML files to my IIS directory.

AppManager does what it's supposed to do—monitor systems and prevent problems from arising. AppManager will meet the needs of any administrator who works with large enterprise networks. However, the product is plagued by its price. At $2500 per console and $600 per server, AppManager is expensive to implement. If all you need is a tool to ensure that your Exchange server's performance is optimal, justifying the cost of this solution will be difficult. However, if you make the initial investment, you can integrate other AppManager network-monitoring modules without breaking the bank. If you anticipate needing a complete enterprise-management solution in the future, you can't go wrong with AppManager.

bv-Control 1.5 for Microsoft Exchange
bv-Control is a useful Exchange Server management tool. This package includes two components: bv-Control, BindView's risk-assessment management tool, and the IntelliPACS console, a realtime network-monitoring facility.

Installing bv-Control is straightforward. After I specified in which directories I wanted the setup program to install bv-Control, the program configured the software and let me select the system components that I wanted bv-Control to monitor—Exchange Server, in my case. I provided the program's licensing information, and bv-Control was ready to run. To install IntelliPACS, I simply launched the setup program and the software automatically installed.

bv-Control uses a client/server architecture. In an ideal environment, the server component, BindView Information Server, runs on a system that contains bv-Control's data-collection facilities, and the client component, BindView Risk Management Solution (RMS) Console, runs on another machine from which you monitor the Exchange Server system. However, you can install both the client and the server components on the same system without affecting the system's performance. I installed both components on my Exchange Server system.

BindView designed bv-Control as an MMC snap-in, and you can access BindView RMS and IntelliPACS from the MMC framework. However, the software has a learning curve. Although the UI clearly labels the software's features, configuring the software can be unwieldy. Almost every operation requires multiple configuration steps, and I had to crack the manual on many occasions, even for tasks as simple as defining a policy.

To configure bv-Control policies, you use queries, which are a set of structured questions that interrogate your Exchange Server system to obtain the information you want. Creating new queries with bv-Control is a fairly simple process: You click the New Query icon on the toolbar and select the appropriate data source or operation type you want to query for. Then, the software prompts you to select the appropriate data fields to add to the query. After you select the data fields, you can set bv-Control to filter out specific events. In addition, you can narrow your query to a specific group of servers or mailboxes. The query builder's event-based approach comes in handy for administrators who don't relish the idea of writing code to create custom policies.

bv-Control also offers many predefined policies, which the software conveniently groups by policy type, as Figure 2 shows. BindView's predefined policies covered every task I wanted to automate (e.g., search message stores to find all email messages that contain a specific topic, create a list of orphaned email accounts), so I didn't need to create custom queries.

To test bv-Control's querying feature, I used the predefined Traffic Analysis queries to create a grid view of the amount of data my Exchange Server system was handling. To create an organized chart that showed the users on my Exchange Server system and the amount of disk space they consume, I ran the software's Mailbox Disk Space Utilization Report query. After I created both reports, I manually exported the data and saved the reports as HTML documents to publish on my IIS system.

The IntelliPACS component handles all system monitoring. At first launch, IntelliPACS prompted me to designate a computer on the domain as the Event Database Server (EDBS). IntelliPACS uses SQL Server to store information, so I designated my SQL Server system as the EDBS. In addition to this information, IntelliPACS prompted me for a system account and the default systems administrator account on the SQL Server machine.

IntelliPACS uses an agent-based design to monitor applications and services. You deploy an agent on each computer that the main IntelliPACS server will monitor. Each agent can run multiple scripts, which the software uses to monitor specific aspects of each system. By default, IntelliPACS includes many predefined scripts, and you can easily use VBScript to create custom scripts. Alternatively, you can use IntelliPACS's Script Wizard to create custom scripts. I chose to use the wizard to create a script to monitor the state of my Exchange Server system.

Before you use the wizard to develop custom scripts, you must define the alert types you want IntelliPACS to use in the event of a system error. I set up IntelliPACS to issue a Net Send command to my workstation and create an entry in the event logs when an error occurs. If you require more aggressive alert actions, you can set up IntelliPACS to send pager messages, email, and SNMP traps, or to launch an application.

After I set up an alert, I launched Script Wizard and created a script called Service Status. I selected Windows System as the script category and NT Services as the resource data. I set the script to poll my Exchange Server services every 2 minutes. I planned to use the information that the script gathered to create a historical trend chart that documented the frequency of my Exchange Server's crashes, so I enabled the software's data-logging feature. Finally, I selected the services that I wanted the script to monitor and set IntelliPACS to automatically restart the services if they failed. I then manually killed every Exchange Server process. Within 2 minutes, IntelliPACS sent a warning to my workstation and restarted the services.

For Exchange Server management, there's not much that bv-Control can't do. Although the UI can be clumsy, the program's well-written documentation simplifies finding the information you need. The combination of bv-Control and IntelliPACS makes server management and monitoring easy. However, at $1995 for the console, $995 per monitored server, and $12.95 per monitored mailbox, you'll be paying as much for an Exchange Server management utility as you would spend on a general-purpose management package for your whole system.

Foglight 3.0 for Exchange
After acquiring MessageWise, Quest refined MessageWise's 2MA package by adding template support, HTML-based reporting features, and a new architecture. The result was Foglight 3.0, a dedicated Exchange Server management and monitoring tool.

Installing Foglight is easy; you simply run the setup program on the computer you designate as the Foglight server. The software executes the setup process in two phases. The first step installs the program, and the second stage installs and configures the service. After the system completes the first step and reboots, you must leave the CD-ROM in your CD-ROM drive for the software to complete the second step. The documentation was unclear about this requirement.

After installing Foglight, I launched the console, which isn't an MMC snap-in, and created a zone for my domain. In Foglight parlance, zones are a collection of servers that the software monitors as a group. You can use the Add Server Wizard to quickly add and configure the servers in your zone. If your network contains multiple Exchange Server systems that service different machines, you can create multiple zones and manage them from a central console. After using the wizard to configure the first server in the zone, I opted to manually install each additional server—a process that was almost as simple as running the wizard. To manually enter my Exchange Server system into a Foglight zone, I entered the server name and selected the appropriate check boxes to designate the Exchange Server services for the software to monitor.

I have a designated SQL Server system, so I decided to pass on the Jet database engine that Foglight uses by default and created a SQL database instead. The difference between the two database formats is size. If you plan to collect a large amount of data or store your collected information to create historical trend analysis charts, use SQL as the back end. Otherwise, the Jet engine is adequate. After I supplied Foglight with a SQL Server account, the software was ready to go.

Foglight uses what Quest calls a Touchless Architecture, which means that Foglight doesn't rely on agent software. Instead, the program runs on a central server and uses the remote procedure call (RPC) protocol to collect information from client machines. This design simplifies the deployment process because you don't have to install additional software on the client systems. In addition, this architecture doesn't impose the amount of overhead that agent programs do. Furthermore, this overhead reduction doesn't come at the expense of functionality. By using RPC to monitor services on remote systems, Foglight can collect data that the remote system's Performance Monitor counters provide. The drawback of this architecture is that if a system loses its network connection, Foglight stops collecting data, whereas an agent-based architecture will continue farming data to send to the central repository when the remote system reestablishes a network connection.

Template sets define Foglight's policies. Template sets are groups of rules that the software categorizes by service type and that you can apply locally or globally. Thus, you have all the power and customizability of rules without hand coding. You use the software's console, which Figure 3, page 68, shows, to configure templates by defining events and conditions. For example, to create a template that tracks traffic on my Exchange Server system, I selected the Average Time For Delivery parameter from the template list, set the threshold counter to trigger an alert when message delivery falls below a specified value, defined a sustained value that tells Foglight when to send a notification message, and saved the template. Then, I applied the template to my Exchange Server system's zone, and the software immediately began monitoring. Quest includes many templates, all of which correspond to the appropriate Performance Monitor counters, so you probably won't need to create templates.

After configuring the necessary templates, I tested Foglight's monitoring capabilities by simulating three events that frequently arise in a live environment. First, I created a large dummy file to fill up all the available disk space on my Exchange Server system. I had set Foglight to send me an email message after the software checked the server's disk space three times to confirm that the lack of disk space was permanent. So, the program sent me an email message. Second, I created a bottleneck between the Exchange Server machine and the network. Again, Foglight noticed the slow connection and sent me a notification message. Finally, I killed the Exchange Server services. The program perceived that it wasn't receiving information from the services and notified me of the downed server. Unfortunately, Foglight doesn't automatically restart failed services.

Foglight takes three approaches to reporting. The first approach is email reporting. When you select this option, Foglight emails text-based reports to you at regular intervals. The second option is in-console reporting, which lets you select a reporting period from within the Foglight console and output a chart-based report. The third option, Web reports, is the most fun. The Web reports approach lets you use HTML to create metric reports so that you can use any Web browser to view the reports.

At $1195 per Foglight server, deploying Foglight in your enterprise won't bust your budget. The cost is all-inclusive; thus, you don't pay additional fees for the console and per-user licensing agreements. You can even install Foglight on multiple machines without paying an additional licensing fee. The software is easy to use, so you won't have to learn confusing interfaces or scripting programming languages. If all you require is simple Exchange Server management and monitoring, I highly recommend Foglight.

HP OpenView ManageX SMART Plug-In 4.0 for Microsoft Exchange Server
Designed as an MMC snap-in, ManageX is a systems management tool designed to monitor diverse IT environments. HP and Microsoft partnered to develop ManageX, which relies heavily on standard Windows technology such as Windows Management Instrumentation (WMI), Distributed COM (DCOM), and VBScript. ManageX acts like a part of Windows.

At its base level, ManageX monitors essential services such as Win2K Server Terminal Services, Microsoft Transaction Services (MTS), IIS, and Systems Management Server (SMS). For more comprehensive coverage beyond basic health and status monitoring, HP ships application-specific modules called SMART Plug-Ins. Each module provides additional in-depth functionality, such as performance tuning and enhanced policies.

The installation process was uneventful: After I ran the setup program on my Exchange Server system, the ManageX plugin for Exchange Server (the setup program installs ManageX and the plug- in) took up about 175MB of disk space on my hard disk and didn't require a reboot. The setup program also installs product documentation in Adobe Systems' Adobe Acrobat format. The documentation includes a useful tutorial that eases the software's learning curve.

Boasting a powerful auto-discovery feature, the ManageX plugin immediately detected my PDC, IIS, Terminal Services, SQL Server, and Exchange Server systems, as well as the clients connected to my domains. Navigating the UI is simple, especially if you've been working with Win2K's UI. Because the ManageX plugin snaps in to the MMC, you'll probably be familiar with the software's nuances. However, HP went one step beyond an MMC snap-in and gave the software a Web-based UI, so you can view the UI from Web browsers on remote systems.

After the ManageX plugin probed my network configuration and created an organizational map that included a visual message flow of my systems, I began configuring policies by drilling down through the console to the appropriate policy, as Figure 4 shows. The software lets you define policies as expressions or VBScript code. Expressions are abstract strings of constants and variables, so expect to spend some time with the manuals to master this capability. If you have a rudimentary grasp of BASIC, you can avoid using the ManageX plugin's expression syntax by using VBScript to create custom policies. The software also offers a healthy list of prefabricated policies that you can fit into almost any configuration type. I opted not to touch the default policies and disabled the ones that were unnecessary in my environment.

For performance reasons, I have multiple servers running my core services. Rather than manually configure policies on each server, I used the ManageX plugin's alias definition feature to push my modified policy list to servers on an as-needed basis, which I defined by giving the software criteria to filter with. I have Terminal Services and IIS on Win2K systems and SQL Server and Exchange Server on NT 4.0 machines. By telling ManageX to deploy Terminal Services and IIS policies to the Win2K domain controllers (DCs) and apply SQL Server and Exchange Server policies to the NT 4.0 machines, I set up a customized monitoring solution.

To test the solution's monitoring capabilities, I simulated several events on the Exchange Server system. First, I created a large dummy file to fill up the spool disk. Not surprisingly, the ManageX plugin sprang into action and prompted me with a pop-up message that notified me of the low disk space on the server. I then created a bottleneck on the network by sending multiple gigabytes of data back and forth between two segments on the switch. The software's threshold counter detected this aberrant behavior and notified me that congestion was compromising mail delivery. Finally, I physically disconnected the Exchange Server system from the switch. The ManageX plugin immediately noticed the missing server and notified me through a pop-up message.

The product's report generator, Smart Reporter, takes an open approach to statistical analysis and lets you fine-tune the amount of information that the generator analyzes. Smart Reporter's engine uses the preconfigured logging policies to farm data from each machine. The central server then collects the data and stores it using either Microsoft Access or SQL Server. This setup lets the software create reports based on the templates you select.

To monitor traffic on my Exchange Server system, I modified a summary detail template and set Smart Reporter to grab a sample week's worth of usage statistics. A few seconds later, the ManageX plugin presented a top-level view of the data that was neatly analyzed, correlated, and ready to be exported to my IIS server so that I could maintain a historical trend analysis of my server's usage. I then created a summary report to show the amount of space each user was using. After running the report generator, the ManageX plugin gave me a point-in-time list of the amount of disk space each user was using as well as the number of messages in each mailbox.

Because HP based the ManageX plugin's UI on the Microsoft Internet Explorer (IE) HTML renderer, the software exports all generated reports as HTML files native to the console. To automatically publish reports to an IIS server on the network, enable the Publishing feature, which lets you view reports from any machine on the network on which a Web browser is installed. Alternatively, you can configure the ManageX plugin to generate reports in Word format.

ManageX is an excellent enterprise-management tool. The only drawback is the solution's cost: $2995 for the ManageX console, $795 for the plugin, and $1195 per server. If your only concern is Exchange Server monitoring, you might be paying for features you don't need, and a more application-specific and less expensive option would be better for you. If you purchase ManageX only for the purpose of monitoring Exchange Server, you'll probably eventually find yourself using it as an enterprisewide management tool.

PATROL 3.4 for Microsoft Exchange Server
BMC takes an interesting approach to software distribution. PATROL ships as part of BMC's PATROL family of enterprise-management products, which includes support for almost every back-end system platform. Although this distribution approach might seem fairly standard, the licensing plan you purchase dictates which of the product's monitoring modules you can actually use. When you open the product's box, you discover a slipcase that contains five CD-ROMs and five spiral-bound manuals. A maintenance update that BMC sent out shortly after the release of PATROL adds three more CD-ROMs to the mix.

I installed only PATROL, the Win2K and Exchange Server 5.5 modules, and the PATROL Explorer and PATROL Operations Manager components. The setup process unfolds through three main steps: installing the console, installing the agent, and installing the Knowledge Module Deployment Server (KDMS). After the installation was complete, I installed the PATROL maintenance release to update the software. After the system update was complete, I repeated the agent installation on each client machine. Each agent then uses TCP/IP to communicate with the central server. After I configured the client systems, I added them to the PATROL console, and the software began monitoring the network.

On the server side, BMC bases PATROL's architecture on three components: PATROL, the server component that runs on the back end; PATROL Operations Manager, the suite's event-management facility; and PATROL Explorer, the management console that handles event information. On the back end, PATROL uses SQL Server as a repository for collecting and storing trend data for historical analysis.

After the installation process was complete, I opened PATROL to discover that it's not an MMC snap-in. Instead, PATROL offers two distinct interfaces: the Developer Console and the Operator Console. BMC designed the Developer Console as a superset of the Operator Console. The enhanced console includes support for rule development, UI modification, and remote agent configuration. For testing, I used the Developer Console to configure and maintain PATROL.

The PATROL Developer Console is busy, as Figure 5 shows. The amount of information displayed can make learning how to use the software a daunting task. I often found myself buried under layers of menus and cracking the manuals for help performing simple tasks such as getting the status of the server. Although power users will love the power stashed within this UI, I would welcome a more streamlined interface.

For monitoring, PATROL's Knowledge Modules define rule sets. Knowledge Modules are a group of predefined rules that are tailored for a specific environment. The Knowledge Modules I worked with were all Exchange Server focused. The software subdivides each module into parameters (monitoring options) and commands (syntax that you execute). Although this configuration sounds confusing, you won't have to worry about it because everything is preconfigured.

To test PATROL's monitoring capabilities, I enabled a module to monitor the amount of disk space on the Exchange Server system. After I created a large dummy file to fill up the space on the Exchange Server system, the agent immediately reported the low space conditions to the PATROL server, which in turn notified me of the problem through a pop-up message. I then enabled the performance Knowledge Module and set it to track the message delivery schedule on the Exchange Server system. After I created an artificial bottleneck between the Exchange Server system and the rest of the network, PATROL noticed that Exchange Server was delivering mail at an excruciatingly slow rate and notified me of the problem. Finally, I shut down the Exchange Server services on the server to see how PATROL would react. Surprisingly, the program not only noticed the downed services but restarted them. If your environment depends on mission-critical services, you'll appreciate this feature.

To create a trend analysis chart, I used the Collect Data agent to consolidate historical data from multiple systems. This agent provides a wizard-based interface to let you select the computers from which to farm data, the database on which to store the data, and the date you want to interpret. After the software collects the data, you can run the Visualizer agent against the database to create an organized chart, which you can then export to several file formats, including HTML.

PATROL is a thorough and comprehensive Exchange Server management utility. However, it's also unwieldy and confusing to use. BMC designed the PATROL suite as an industrial-strength network-management program of which the Exchange Server component is only a small portion. The price reflects this idea: $3500 to $5000 per console, $815 per module, and $695 per server. In addition, the product's complex deployment process means that you won't be able to install PATROL over the weekend. Competing packages that are less complicated or less expensive duplicate much of PATROL's feature set; thus, I don't recommend this solution.

RoboMon 7.6 Exchange Intelligence Solution Set
Tailored to monitor Win2K- and NT-based systems, RoboMon is best known for its ability to manage and monitor complex infrastructures. Heralded as one of the most scalable network-management packages on the market, RoboMon has developed a reputation as a flexible and reliable solution.

RoboMon's installation process is straightforward. Run the setup program and select the system components you want the software to monitor, and the system is ready to go. For this review, I installed only rule sets for Win2K and NT, RoboMon Administration, and Exchange Server monitors. This decision was prudent, considering each rule set consumes about 30MB of RAM—a major consideration if your systems are resource-starved. I also ran RoboMon's SQLLogin utility to give the monitoring tool a system account to use when connecting to my SQL Server system. RoboMon also supports the Jet and Oracle database engines.

After the software installed, I opened the RoboMon Enterprise Manager console to begin configuring the software. Rather than installing as an MMC snap-in, RoboMon uses a self-contained UI. RoboMon's two-pane view interface, which Figure 6, page 72, shows, is intuitive and easy to use. RoboMon also offers a Web-based UI, which lets you access the software from any system on the network.

At first launch, RoboMon automatically queried the systems' services to inventory what was installed on the machines in my network. Before I could add other servers on my domain to RoboMon, I had to manually install the agent software on each system. I have only a handful of servers running in my lab, so this requirement was just a minor inconvenience. But in a large-scale environment, having to manually install an agent on each system is a major hassle. If you're working with multiple domains, you can take advantage of RoboMon's Virtual Domain feature, which lets you group servers logically. For example, you might want to create an Exchange Server virtual domain that contains every Exchange Server system in your organization regardless of the actual domains to which the systems belong. After installing an agent on my Exchange Server system, RoboMon added the system to the domain list, automatically configured the rule list, and began monitoring the server.

RoboMon uses VBScript as the basis for its policies, so creating customized scripts is a fairly simple process if you have some knowledge of the BASIC programming language. For administrators who aren't familiar with BASIC or who prefer not to write code, the software provides Rule Designer. This tool offers a wizard-based UI that lets you create custom rules; however, using Rule Designer to do so can be as complex a process as writing scripts. To use Rule Designer to set up a custom rule, you must select the appropriate event type, specify the polling interval for the rule, and set the conditions. Although this process sounds simple, declaring the conditions for the rule is tricky. To set conditions for a new rule, you must create an expression that defines the rule. For example, to set conditions for even a simple connection-analysis rule, you must select the appropriate statistics and show the relationship between them using variables, operators, and functions. However, if you're willing to invest the necessary time to learn the nuances of RoboMon's scripting syntax, you'll be rewarded with a level of flexibility that surpasses most rule generators.

Fortunately, RoboMon includes a comprehensive set of predefined rules. I used the Database_Disks rule to monitor the amount of free disk space on the Exchange Server system. When I created a large dummy file that filled up the disk space on my Exchange Server system, RoboMon immediately notified me of the low space conditions through a pop-up message. I then created a bottleneck between two systems in the network and used the IMC_Failed_Connections rule to log the connection problems. This rule gave me a bird's-eye view of the problem. Finally, I enabled the Exchange_Service_Not_Running rule and purposely halted the Exchange Server services. RoboMon reported the downed server, but doesn't have the capability to restart the services.

RoboMon truly shines through its reporting and graphing utility. You use the RoboMon Report/Graph Manager to analyze the data that the software collects from Performance Monitor counters, log files, and RoboMon's SQL Server repository. I created a traffic report by selecting a month's worth of traffic for the software to analyze. After I clicked Run, the Report/Graph Manager pulled the relevant data from the SQL Server repository and created a graph that detailed the past month's usage level. To create a list of the amount of disk space each user was using, I selected the appropriate template and modified the criteria. The Report/ Graph Manager created a chart of the users on the server, the size of their mailboxes, and the number of messages in each user's Inbox. You can export and save reports and charts as different file formats. I chose HTML so that I could remotely view my reports from a browser.

RoboMon's monitoring and management features are top-notch, and the software supports many applications and services. Unfortunately, RoboMon is marred by its complexity. The software ships with a plethora of preconfigured rules, but the rule editing system is needlessly complex. RoboMon is an all-in-one network-management solution, as the price suggests: $1195 for RoboMon 7.6 Exchange Intelligence Solution Set, and $3880 for RoboMon 7.6 Exchange Intelligence Solution Set with RoboMon Event Monitor and Management Console. If all you need is a utility to monitor and manage Exchange Server, RoboMon is overkill.

In the End
To select the right Exchange Server management and monitoring utility for your environment, you must consider features such as rule set customization, monitoring capabilities, ease-of-use, and price. Spending more often buys growth potential and might be wise if you'll need to monitor other applications in the future. But if all you need is an Exchange Server focused tool, the expense of many solutions will be hard to justify.

With these considerations in mind, I found Foglight to be the best product of this review. Foglight works the way a network-management product is supposed to work: You install it, tweak the predefined rule set to suit your needs, and don't worry about it again until network problems arise. Foglight is the least expensive product of the pack, but it doesn't scrimp on functionality. If you need to monitor and manage your Exchange Server systems, Foglight will do the job admirably.

However, consider that Quest designed Foglight for Exchange Server. If you anticipate requiring more comprehensive coverage in the future, this product won't grow with your needs. If you foresee adding more server-side tools to your environment, an all-inclusive product will serve you better.