Server Rooms Standardize to Simplify

by Drew Robb

Following standards means more than just buying an "industry standard" product. From evolution to acceptance, standards are critical for both vendors and data center staffs. But what exactly are they? And which standards should you be watching?

There is much talk, these days, about emerging standards. Systems Management Architecture for Server Hardware (SMASH), Common Information Model (CIM), Storage Management Initiative Specification (SMI-S), Internet Small Computer Systems Interface (iSCSI) and Serial Attached SCSI (SAS) are just a few of the many out there. They are, however, among the most critical standards for today's server room administrators.

How Standards Evolve

In an ideal world, standards evolve free of any and all vested interest. End users spontaneously get together to discuss technologies and devise standards for the common good.

We do not live in an ideal world, however, and this idyllic vision is unlikely to be fulfilled any time in the near future. The reality is it takes a LOT of time to create a standard. Massive amounts of effort on all fronts are required to evolve a specification, create a consensus among the standards committee, pilot it, iron out the bugs, and then have the market and end-user community accept it. As a result, standards bodies tend to be formed primarily by vendor representatives.

Yes, this often means some items within the standard itself may forward a variety of vendor agendas. But user enterprises act as a sort of "check" on this system by accepting and proliferating a standard or ignoring it. Overall, vendors do a good job in attempting to find common ground that helps the IT industry as a whole move forward.

Take the case of SMI-S. It was developed by a committee operating within the Storage Networking Industry Association (SNIA). Members are akin to a Who's Who of storage vendors. This has led to some criticism that SNIA is mainly looking after its own — SMI-S has reduced development costs among storage vendors, for example. That isn't bad, and any standard would be expected to produce such a benefit, as the associated hardware and software eventually becomes less proprietary and more of a commodity item. Further, SMI-S has produced some benefits for the end user.

Emerging Standards for the Server Room
Standard Acronym
Systems Management Architecture for Server HardwareSMASH
Common Information ModelCIM
Storage Management Initiative SpecificationSMI-S
Internet Small Computer Systems InterfaceiSCSI
Serial Attached SCSISAS

"SMI-S is the only significant storage standardization effort currently in existence," says Mike Karp, an analyst with Enterprise Management Associates. "It has made things easier for storage administrators in terms of discovery."

According to Rob Callaghan, chair of the SNIA's Storage Management Forum (and also senior product manager at ADIC), SMI-S is an interface that enables multiple vendors to supply different types of storage hardware and software products to reliably and seamlessly interoperate so as to better monitor and control resources. SMI-S enables common communications between devices and the applications that manage them.

"SMI-S can help users ease configuration, discovery, provisioning, trending, event management, security and asset management tasks," says Callaghan. "What SMI-S isn't, though, is a way to manage the data that is stored using devices such as arrays, switches, tape libraries or NAS boxes."

Another storage standard worthy of note is iSCSI, a standard ratified by the Internet Engineering Task Force. It allows the use of the SCSI protocol over an IP network. It is an excellent standard for server administrators interested in building a more sophisticated storage environment. In the past, that meant learning a whole new skill set — Fibre Channel (FC) — and with that the purchase of expensive and difficult to manage FC gear. Now, all enterprises need is a knowledge of Ethernet to hook up a basic Storage Area Network (SAN).

"iSCSI SANs are easy to use, have a lower price than FC SANs and have sufficient performance for certain types of applications," says Natalya Yezhkova, an analyst at IDC. "Although many iSCSI storage systems target SMBs, we believe that larger organizations will represent a more immediate opportunity."

SMASH Those Servers

Another body, known as the Distributed Management Task Force (DMTF), has issued a standard known as Systems Management Architecture for Server Hardware (SMASH). Benefits of SMASH include reducing the management burden of server hardware, cutting the cost of server administration, improving the reach of system administrators to remotely located servers and standardizing the management of heterogeneous environments.

"Server management is currently done on a 'per-server' basis while IT services are offered across a constellation of servers. So there is an inherent disconnect between offering IT services and managing them. Standards become a way to tear down that wall." — John Humphreys, IDC analyst

"Server management is currently done on a 'per-server' basis while IT services are offered across a constellation of servers," says John Humphreys, an analyst with IDC. "So there is an inherent disconnect between offering IT services and managing them. Standards become a way to tear down that wall."

SMASH attempts to resolve this by making it much easier to manage servers regardless of the vendor. Its main value comes from its Command Line Protocol (CLP), which provides a way to manage any number of different types of servers with the same simple script. Instead of scripts for every vendor and sometimes for each server, IT managers can use one script for the same function in every server in the data center. Some of the SMASH CLP scripts, in fact, are as short as two words. An earlier DMTF project — the Common Information Model (CIM) — is now built into most existing servers and leveraged by SMASH to simplify scripting.

As well as managing servers as a whole, SMASH addresses individual components. A server, for example, may have multiple processors, sensors, network cards, logical devices and cooling systems. SMASH can be directed at specific processors, components and subcomponents. Therefore, you can set up a script to periodically check the temperature sensors on all machines to see how power and cooling needs can be adjusted during the day or night to avoid overheating of equipment or reduce the electric bill safely.

"System managers told us they wanted one command-line protocol they could write scripts for and interact with in a consistent way," says Winston Bumpus, president of the DMTF. "They were fed up with having to use scripts for each separate vendor. Even across product lines within vendors, the scripts required can be different."

SMASH-based server products began shipping this year. For enterprises deciding between two comparable servers, one of which adhere to SMASH and the other does not, it makes sense to opt for the SMASH model.

Industry Standard?

Another "standard" that cadres of server, chip and OS vendors consistently bandy about is "industry standard." What does this mean? Essentially, they are talking about servers from Dell, HP and IBM; chips from Intel or AMD; and Microsoft software (though some now include Linux within this category). What is called a standard is really just a convenient partnership among these and other vendors to sell their own wares.

Sun Microsystems is perhaps the last of the major OEMs to buy into this propaganda. Its SPARC-based products have been under attack for years from the Wintel camp. Although Sun has been upgrading its SPARC-line and leads the way in multicore chips, some would say it is inadvertently forwarding the campaign against itself.

"Sun has moved into the industry standard space with our servers with the idea of horizontal scale," says Kathleen Holmgren, vice president, disk systems business, Data Management Group, Sun Microsystems.

The question Sun should ask itself is this: If its x64 AMD-based servers are "industry standard" does that not imply that UltraSPARC and PrimePower servers are non-standard?

There is nothing wrong, as such, with buying servers and software from the vendors behind this campaign. But it smacks of the same scent as the 1990's propaganda that steered IT managers away from proven platforms, such as the mainframe and AlphaServer, onto immature Wintel systems that really needed another decade before they could deliver similar levels of reliability and performance.

Now that these systems can perform relatively well in the enterprise, it appears they are gearing up to knock out the rest of the competition — RISC, mainframe, OpenVMS and Unix. If they succeed, the industry standard will become one by virtue of the fact that no other alternatives remain.

This article was originally published on Thursday Oct 19th 2006
Mobile Site | Full Site