by Steve Rokov
Previous Guest Columns

Trust the chip advantage
by Steven Sprague
July 2005

How to manage telecom expenses
by David C. Perdue
June 2005

VoIP for the SMB
by Dan Murray
May 2005

Manage your network security
by Carl Herberger
April 2005

Hosted telephony pays off
by Alaric Silviera
March 2005

Simplify your distributed network
by Doron Abrahami
February 2005

Leave it to the experts
by Chuck Machlin
January 2005

Emerging wireless: Who’s on first?
by Chris Couper and Marilyn Murphy
December 2004

Collapse of the ‘Web tier’
by Craig Stouffer
November 2004

Service-continuity goals important
by Malcolm Fry
October 2004

Trends in WAN outsourcing
by Vab Goel
September 2004

The patching game
by Eric Vasbinder
August 2004

Policy-based networks: Why not further along?
by Steve Pettit
July 2004

Solve the bandwidth dilemma
by Teejay Riedl
June 2004


Identify your storage options
by Paul Mayer
May 2004

Visualize the virtual network
by James Leach
April 2004

Maximize the power of fax
by Tom Linhard
March 2004

Who will dominate Web conferencing?
by Ian Widger
February 2004

NAS gains traction
by
Joe Disher
January 2004


Make blades easier to manage

Try applying some “glue” to facilitate less painful management of your blade servers.

The promise of blades has not yet been fulfilled. Too many proprietary components and management tools have created added complexity, resulting in higher operating costs. Enterprise IT managers are looking for blades to be as manageable, accessible and interoperable as traditional servers, while offering the space- and power-saving advantages inherent in blades. The common “glue” that today binds many traditional servers is what enterprises also require from blades.

Blade glue is the standardization of silicon, firmware, drivers, providers, utilities and applications that provide blade managers with always-available solutions. Unlike proprietary approaches, blade glue provides standard management capabilities that are common across multiple platforms, making management of blade systems less painful.

There are three layers to blade glue:

  1. Foundation. Pre-integrated components built on industry standards; crash proof and tamper resistant; hardware-level health monitoring and access control; autonomous systems monitoring/control interface throughout all system states; and standardized command line.

  2. Infrastructure. Virtual presence–keyboard, video, mouse (KVM) redirection; virtual media; and modular systems management.

  3. Software interoperability. Use existing management tools and extensible with additional technologies and standards.

Some of the key technologies and standards that provide these features include:

Intelligent platform management interface (IPMI). This specification represents a proven approach to hardware health monitoring and management. IPMI is typically pre-integrated into a blade and chassis by the vendor.

Specialized IPMI software running on dedicated controllers allows IPMI to act as a completely separate subsystem immune to the condition of a blade. This independence allows administrators to utilize IPMI features any time via IPMI-enabled remote applications. Monitoring the health of components, such as temperature, power and inventory information, can be done even during times where the blade operating system may have hung or when the blade is inactive. IPMI also features user-defined thresholds for specific events.

Systems-management architecture for server hardware (SMASH). This is a Distributed Management Task Force initiative that is delivering, among other things, a standard command-line protocol (CLP) to provide a consistent way to manage servers from multiple vendors. The first version was expected to be available for public review mid-2005. The goal of CLP is to offer a standard way for administrators to send commands and receive responses back from managed servers.

Think of SMASH as the primary interface, out of the box, accessing IPMI features in the box. This command consistency across different vendors’ servers or blades is critical, not just for reducing the learning curve, but especially when considering investments in scripts that many customers have made. CLP typically will be supported on the blade in firmware and elsewhere in the chassis via management modules–offering multiple access and control points.

Keyboard, video, mouse. By redirecting the KVM from multiple servers to a single console, considerable cost savings in space and equipment are achieved. As a hardware approach, it does not require any additional software on the blade server, and is typically accessible over the LAN from remote locations.

Using IPMI features, an enterprise manager can learn that the temperature of a blade has exceeded a threshold, making him aware of a potential failure. With SMASH then serving as a common command interface and a remote KVM appliance offering cost-effective centralized, graphical access, the manager can remotely and gracefully power down the blade and activate a backup to keep services running while scheduling the blade fan for replacement.

For more information from Avocent:
www.rsleads.com/508cn-258

Steve Rokov is director of technical marketing, Avocent Manageability Solutions Group. He is also the secretary of the Blade Systems Alliance, an industry consortium seeking to advance the adoption of blade-based computing systems. Send comments for publication to guest@comnews.com.

This is the final installment of our Guest Column series. Beginning in September, a new column, IT Insight, featuring an interview with an enterprise IT manager, will take its place.