ChipCenter Questlink
SEARCH CHIPCENTER
Search Type:
Search for:




Knowledge Centers
Product Reviews
Data Sheets
Guides & Experts
News
International
Ask Us
Circuit Cellar Online
App Notes
NetSeminars
Careers
Resources
FAQ
EE Times Network
Electronics Group Sites

DESIGNING A DSP-BASED RAS SERVER


Circuit Cellar Online
THE MAGAZINE FOR COMPUTER APPLICATIONS
Circuit Cellar Online offers articles illustrating creative solutions
and unique applications through complete projects, practical
tutorials, and useful design techniques.

DESIGNING A DSP-BASED RAS SERVER

Lessons from the Trenches Part 1: RAS Server Background
by Shawn Arnold

Start ý The Little Picture ý RAS Port ý Host Controller Service Details ý RAS Port Design ý Sources and PDF

RAS PORT DESIGN

Now that I've examined the major blocks of a RAS server from a functional perspective/standpoint, it's now time to examine some of the critical design goals and design criteria of a RAS server. Our focus is on the RAS port block of the RAS server, so Iýll focus on the critical specification and important features of the device that will implement the RAS port portion of the server.

As a manufacturer of networking equipment or, more specifically, RAS servers, in order to make money, I have to offer "more" in the final product, but at the same time, I must build it with fewer resources! In the case of a RAS server the most important "more" is ports.

The critical design goal becomes maximum port density (i.e., I want the RAS server to contain the greatest number of ports in the smallest amount of space). RAS server manufacturers are pushing for maximum port density, and thus, for a greater number of ports on each new generation of RAS server products.

Coupled with the desire to reduce port size, is also the desire to minimize the power consumption of each port. Even though the next generation RAS server product will have a greater number of RAS ports in the same amount of space, the manufacturer does not want an increase in the power-consumption specifications of their product. Increasing the power-consumption specifications may cause serious difficulty for their customers, who may have to redesign power systems.

FLEXIBILITY

Along with lowering the cost of each port by reducing physical size and lowering power consumption, I also have to offer more features within each port. As briefly mentioned earlier, the port must be able to handle the various type of RAS services such as data, fax, and voice.

Ultimately, I would like to have these various services supported in a dynamic sense. I donýt want the restriction of having to dedicate certain ports to certain services (i.e., I do not want the service built into the port). Rather, I'd like each port to be reprogrammable or reconfigurable in real time so I can configure the port on demand to support the service required by the remote or network user. Such a real-time configurable port solution is referred to as a universal port.

COMPLETENESS

In today's highly competitive market place, the shortest time to market is paramount. This means that the design cycle must be as short as possible.

Designing a RAS server not only includes the physical construction of the hardware that implements the physical system, but it also includes designing the software that must be developed to orchestrate and breath life into the hardware. This software includes code that runs on the various controllers and data processors in the RAS server, including RAS port firmware for modem, fax, and voice services; host controller code for port and network I/F services; and trunk line transceiver code for the controller code.

Any portion of the above software that can be supplied as part of a product offering is of great value. Many RAS port devices are offered with firmware that implements all aspects of each RAS port service. This is of great value to a RAS server manufacturer because it significantly shortens their design cycle.

The implementation of modem, fax, and voice services is a complicated task. The technology required to implement and test these communication standards is difficult to acquire. Thus, it is highly desirable for a RAS server manufacturer to purchase or license this technology in a ready-made firmware form that will run on the platform chosen to implement their RAS server port. A universal port that includes this technology as part of the total solution is quite desirable.

In a smaller sense, procurement of host code that interfaces to the RAS port can also help to reduce costs and time to market. This is more difficult to find, and in many cases, the host code portion of a RAS server is where the differentiation lies between one manufacturer's RAS server and another's.

So, in this brief introduction, I've discussed the RAS server's hardware block diagram and presented several important design criteria. I covered the three main blocks of the RAS server and the fact that the RAS Port is the heart of the system. Stay tuned for Part 2 where I'll narrow the focus of the discussion and use a DSP to implement the RAS port. You'll see why a DSP is a great choice for a Port device and I'll discuss, in detail, the hardware and software design issues.

PREVIOUSNEXT


Circuit Cellar provides up-to-date information for engineers. Visit www.circuitcellar.com for more information and additional articles.
For subscription information, call (860) 875-2199, subscribe@circuitcellar.com or subscribe online. ýCircuit Cellar, the Magazine for Computer Applications. Posted with permission.
Click here to get your listing up.

Copyright © 2003 ChipCenter-QuestLink
About ChipCenter-Questlink  Contact Us  Privacy Statement   Advertising Information  FAQ