Looking Toward Green Blades of Servers - TvTechnology

Looking Toward Green Blades of Servers

Server- and workstation-based subsystems are a natural course of evolution for what is extensively a software-based system architecture.
Author:
Publish date:
Image placeholder title

We seem to be experiencing an unending proliferation of discrete, independent and separate computer boxes that must be integrated into the technical and production sectors of the facility. The question looms "do we really need all these servers and PC-workstations in our facility?"

PCs have become both the human and signal interfaces for graphics systems, automation, signaling and command/control. In most systems there are PC-platforms for transmission and encoding operations. A plethora of screens and terminals for setup and diagnostic processes are in every system. Next, add in servers for cached content from program providers, coupled with servers for wrapping or transcoding, and proxy servers that are outboard of the play-to-air and ingest servers. Well, you get the idea; it's a seemingly unending extension of devices that are guaranteed to keep the IT-staff employed for an eternity.

ALTERNATIVES

Server- and workstation-based subsystems are a natural course of evolution for what is extensively a software-based system architecture. With operational cost management becoming more prevalent in broadcast and media operations, perhaps the entire industry should take a look at alternatives to the dozens of these individual 200 to 750 watt devices that make up the facility. As part of a focused view on the "green is good" perspective, there may actually be alternatives for broadcast systems that should be considered. The data and telecom industries are each looking at cost control and system management, why shouldn't we?

An internal audit of the equipment deployed and the number of discrete PC-based subsystems might reveal some interesting discoveries. In the plant, some workstation platforms only provide occasional services to the everyday functions. Others, such as transcoding servers or automation may be fully dedicated services providing reliable, mission critical services 24x7x365. Somewhere in the middle are utility systems, elements that provide operational or information interchange between business systems or secondary devices. And still others may simply be there "in-case" we need them or because the software systems knew of no other way to get the isolation or functions needed.

Consolidation of server and PC-like workstations may offer material cost advantages to both capital (cap-ex) and operational (op-ex) expenses. As an example, gateway devices could be a prime candidate for consolidation of discrete systems. The integration of functions onto common platforms may be another.

BLADE SERVER

When the facility finds there are areas where dedicated PC-workstations provide only a single or very narrow set of the processor applications, putting those applications onto slot-based computer cards, such as in a blade server chassis, might very well be an option. A blade server is a chassis housing multiple thin, modular electronic circuit boards, known as server blades. Each blade is often dedicated to a single application. Blades are literally servers on a card, containing processors, memory and integrated network controllers. Some blades have optional fiber channel host bus adaptord (HBA) and other input/output (IO) ports.

Blade server or PC-card elements could offer potential cost savings compared to installing several sets of dedicated PCs each with superfluous features such as comports, graphics drivers and individual hard drives. The blade server chassis shares a common, usually redundant, power supply system. As a result, the footprint in the rack room becomes smaller, and the power and cooling requirements may be significantly reduced.

Small form factor PC-cards allow onboard memory to be reduced and use an operating system that is scaled back to provide only certain needed services. In the cases where full up PC-workstations are required, you could select a streamlined browser as the user interface, and turn off the multitude of unnecessary services sometimes found in PC-based workstations.

The professional broadcast server manufacturers may consider extending this concept in another direction. If the collective volume of services or functions provided by these PC-workstations could be brought into the video server domain, e.g., the transcode process or the proxy generation process, this could potentially reduce the number of physical chassis and points of failure for the facility. Already some of these functions are being integrated into ancillary systems provided by broadcast server, storage and automation vendors. Multiprocessor and multiframe subsystems that offer media-aware extended storage and/or processing are already handling some transcoding or proxy generating functions. Other video server engines are offering options that include multiviewer visual monitoring at their terminal ports, or are added branding capabilities to their baseband output ports.

REDUCING YOUR FOOTPRINT

The future may see these ancillary services become part of a nonbaseband, all file-based systemization. It is not beyond reality that the output with the program service will include features that are integrated only at the file level. Minimizing baseband essence is particularly advantageous to the delivery of content for streaming applications, for transport to alternative sites (e.g., a disaster recovery location), or for program alternative uses such as mobile platforms. Anywhere you can minimize unnecessary decoding or secondary processing, including the physical ports for PCM audio and SDI-video, will reduce manufacturing costs as well as reduce power consumption.

Finally, in the quest for a total solution on a smaller footprint, quality control processing at the file level which is employed directly on the storage platform makes practical sense. While these functions generally require considerable compute-horsepower, having multithread processing capabilities built into the same platform as the ingest-encode and preparation subsystems would allow processor-controlled time sharing that increases operational efficiencies system wide.

It may be a few years before these concepts will find their way into the broadcast or media distribution facility. In the same sense, it could take several generations of user-requirement refinements to narrow down the ballooning number of compression and media formats that we have today. Eventually the industry will find ways to scale the operating facility so that the physical equipment count is reduced while the workflow is streamlined.

In conclusion, these concepts could be employed today through clever design and an insistence by users to establish a more thoughtful integration of features, functions and components. Stay tuned!

Karl Paulsen is chief technology officer for AZCAR Technologies. Karl is a SMPTE fellow and an SBE Life Certified professional broadcast engineer. Contact him at karl.paulsen@azcar.com.