IT managers face increasing complexity in the data center.
As IT organizations expand into multiple server platforms across diverse departments, sites, and geographies, their server management strategy will determine their success in terms of functionality and operating expense.
Associated cost control is a major concern because an IT organization’s largest expense is typically related to ongoing network and systems management.
Not long ago, administrators had limited options. There were no comprehensive cross-platform standards that addressed their need to directly manage servers from multiple vendors.
This led hardware manufacturers to develop assorted tool sets to manage systems via in-band and out-of-band connections for different operating systems and system states.
Today, enterprise server management in many data centers is comprised of a variety of similar tools and applications for each vendor platform deployed. In many cases, administrators have customized the original management tools into ones more specialized for their unique environment, installation, and product.
The resulting array of management commands and tools that differ by provider can be extremely inefficient — and therefore, expensive — in terms of ongoing operational costs.
Open Standards for Server Systems Management
The DMTF created the Systems Management Architecture for Server Hardware (SMASH) Initiative to help address the problems cross platform server management.
SMASH includes a suite of specifications that deliver architectural semantics, standard industry protocols, and profiles to unify the management of the data center and increase productivity.
SMASH specifies both standard command line protocol and web services interfaces.
Administrators often need to interactively manage various systems using a specific command. But servers from different vendors often require different commands for completing the same task.
The SMASH command line protocol provides an interface to heterogeneous servers independent of machine state, operating system or OS state, system topology or access method. It is a standard method for local and remote management of server hardware using out-of-band communication.
The SMASH web services interface provides a standard way for system management vendors to support diversified server management in their tools and implementations.
By offering consistent server management that handles multi-vendor server monitoring and management tasks, managers can improve interoperability, increase efficiency and control operational expenses.
Benefits of SMASH for Administrators
By deploying servers and management solutions that support SMASH server management standards, administrators can manage their multi-vendor server environments in a consistent manner.
SMASH offers flexibility in that it supports a variety of server platforms — from stand-alone servers, to blades, to racks. It is also supports management independent of machine or OS state.
SMASH includes both a command line protocol and web services protocol. This means that administrators can manage their servers for ad hoc tasks using a consistent command set. Or they can write one script that runs against multiple vendor servers.
The web services protocol allows management tool vendors to easily support management of various types of servers in their consoles and tools.
Thus, the administrator can consolidate management of their diverse server environment into a single management console.
Systems Management at the Hardware Level
Platform hardware management generally refers to the remote monitoring of platform hardware variables such as fan speed, voltages, CPU and enclosure temperatures along with a wide range of other sensors.
It also implies the ability to remotely control the power state of the platform and to reset the system back into an operational state should it "hang.”
Despite the advances made in standardizing manageability firmware and software interfaces, many manageability hardware subsystems are still proprietary and are often soldered to the baseboard.
This can make manageability hardware subsystems inflexible in terms of upgrade path and a cost liability for IT managers.
Open Platform Management Architecture
AMD led the development of the Open Platform Management Architecture (OPMA) specification (PDF 1.4MB) to address these issues.
OPMA is an open, royalty free standard for connecting a modular, platform hardware management subsystem (an "mCard") to a computer motherboard.
A significant advantage of OPMA over previous generation management subsystem attachment methods is that OPMA does not consume a PCI socket. OPMA cards are also smaller and cheaper than other solutions.
The OPMA specification covers the following:
- Signal list
- Connector and pin out
- Power requirements
- Mechanical form factor
- BIOS and management controller firmware interfaces
- Division of management subsystem resources between motherboard and mCard.
OPMA enables a wide variety of mCards to individually interface to a given motherboard. It also enables a single mCard to individually interface to multiple motherboard models.
OPMA can support multiple management standards and implementations to deliver various levels of manageability required by the customer. For example, an OPMA card can support server standards like IPMI or SMASH.
OPMA was created as a joint technology development effort between AMD and platform management subsystem technology companies such as Agilent, AMI, Avocent, and Raritan Embedded Solutions (formerly called Peppercon).
Learn more about systems management:
