Another standards front for management is SNMPconf (configuration management with SNMP) (see "A Key Piece of the Management Puzzle"). This best-practices and policy MIB leverages the existing SNMP framework for configuration.

Advantages over the ASCII approach are the consistency already defined by SNMP and the value provided by an interface that combines monitoring and configuration. After all, this was the original vision for SNMP.

Now, however, SNMPv3 fixes this with strong security. It's still a wait-and-see effort, but some industry uptake is beginning. Cisco since IOS 12 has shipped SNMPv3 support, and the DOCSIS cable-modem specifications require SNMPv3 for configuration.

Fault, Performance and Marshmallows

Some think it took mini marshmallows to bring green Jell-O to the pinnacle of epicurean delight; sticking fault and performance together is the latest recipe for mouth-watering application management. And when it comes to making applications deliver business value, enterprises are trying hard to get a firm grasp of the "hows" and "whys."

It's no longer enough just to monitor the network, server or database. Instead, the combined service these solutions deliver needs to be managed (see "Application Performance Measurement Grows Up).

Micromuse's Netcool and Concord Communications' NetHealth, historical leaders in fault management and performance management, respectively, have taken a page from each other and are creating service-assurance products that combine fault monitoring and performance monitoring.

Micromuse has added synthetic transactions and network polling to its Netcool suite of alarm and trap reduction tools. And Concord's NetHealth, able to gather data from anything that will stand still long enough to be polled twice, has added alert and trap gathering with the aim of correlating end-to-end data.

Not to be outdone, the veterans of network and systems management, Hewlett-Packard and IBM, have made their own advances. HP has added a Java console, SNMP OID-based polling, Cisco Discovery Protocol and Layer 2 discovery to its OpenView. What's more, OpenView's HTML reporting of historical performance is stellar.

IBM has enhanced Tivoli NetView's operator-intuitive front end by improving router placement in version 7.1. The actual hierarchy of the network is now framed in relationship to the backbone, and the Web console security supports finer-grained access control.

Also, the ability to identify and represent Multiprotocol Label Switching routers makes NetView more useful in service-provider environments. NetView 7.1 also includes support for IBM's CNAT (comprehensive NAT) routers.

This integration into NetView lets service providers manage with SNMP-networked devices that have static, NAT-mapped, nonunique addresses. NAT networks, a common service provider and enterprise approach for combining IP address name spaces, makes SNMP management impossible. CNAT is the best workaround, and NetView and CNAT are tightly integrated.

Peregrine Systems, manufacturer of the first network-management appliance (see "Peregrine Perches Atop the Pack"), is getting some competition from Silverback Technologies in this arena. Silverback, which initially offered only a managed service, has added a network-management appliance with a Java console, just like Peregrine's. Meanwhile, Peregrine's InfraTools Network Discovery appliance can now manage multiple appliances from a single console.

Have You Been Server-Monitored?

Services continue to be a key piece of the network-management solution, and Web performance services have improved dramatically in the past year.

At the top of this market are Web monitoring pioneer Keynote Systems and veteran application-loading expert Mercury Interactive. Keynote and Mercury have pushed each other and the entire Web performance segment to forge service-based network management (see "Performance Monitors at Your Service").

Keynote's products have defined the Web performance service segment by monitoring e-commerce, marketing the company's crack KeyReadiness Web-loading service, and analyzing traffic to create loads that reflect real-world user behavior.

Not to be outdone, Mercury Interactive has transitioned its expertise in enterprise load testing to a Web service. Mercury's scientific approach to Web load testing is the most thorough. Each step of the process is executed expertly, and you're left with no doubt as to the cause of problems.

In a competitive market that has continued to grow despite the dot-com failures, Keynote and Mercury have leapfrogged each other.

For its part, Keynote has added an advanced Web-transaction monitoring service and has partnered with Adlex, which uses BGP (Border Gateway Protocol), to track Internet performance, which is good for service providers and dual-homed enterprises. Meantime, Mercury now offers a gutsy written guarantee to improve performance, and recently integrated with Freshwater Software to produce a combined system- and user-performance correlation service.

On the desktop-management front, Tally Systems broke ground in May 2001 by offering a service to scan and report on desktop assets (see "Tally Systems' WebCensus: Online Inventory That's a Click Away"). This browser-based inventory service eliminates the hassle of implementing desktop-management suites. Tally was the first--and, as far as we know, the only--desktop-management vendor to transition products to services.

Also interesting for its unique desktop-management approach is the combination of ClearCube Technology's rackmounted desktop scheme and Altiris's desktop-management software (see "ClearCube's Uptime Management Tools Banish the Demons of the Desktop,", and "Is There Such a Thing as a Painless Move? AXMT Says 'Yes'"). The only footprint on a user's desk is a small brick for peripherals and USB I/O. If a user needs a new desktop, a hot spare in the rack is immediately available. And it's all managed remotely. PCs are tucked safely in a rack out of users' reach. You can change or repair desktops in short order, without ever leaving your seat.

WildPackets' software-based EtherPeek NX protocol analyzer is an inexpensive tool that is big on function. EtherPeek has long been known for excellent decodes, and the NX version introduced over the past year offers added performance, IP networking utilities and expert analysis.

In a not-so-small way, Fluke Networks' NetTool has moved portable cable testing ahead. This near-PIM-size tester provides most of what's needed to check a twisted-pair wiring plant. It's quick and accurate, and it eliminates bad cables from the diagnostic mix. If Fluke would only offer it with an organizer and cell service that would reach the basement, it would be almost as good as adding mandarin orange slices and mini marshmallows to your green Jell-O salad.

Bruce Boardman is executive editor of Network Computing, testing and writing on network systems and management. He has 12 years of IT experience managing networks and distributed computing for a financial service provider. Send your comments on this article to him at bboardman@nwc.com.