Communicating Application Specification

by Lloyd Borrett
Technical Cornucopia, May–June 1989

Traditionally, some barriers to effective PC communications have existed between applications and communications software. In contrast to the ease of sending a document to the printer, users who wanted to share a document or data with others had to exit their applications, and call up communications software. Then they encountered a different set of problems.

Users need to conduct their business with clients, suppliers, managers, peers, remote employees, consultants and others. These various groups make up a user's "individual network".

Paper systems — memoranda, letters and air mail shipments — are generally accessible to everyone, but they are slow by electronic standards and do not provide information in a form that is easily manipulated by the recipient.

Electronic delivery systems — modems, LANs, E-Mail and telex — can provide near instantaneous communication, but individual network members will seldom have access to the same electronic communications tool. LANs are the most restrictive since they require hard-wiring of all network members. But even 75 percent of businesses still lack PCs.

A better way

In 1987 Intel began to envision the integration of applications software with communications hardware and software for unprecedented ease of use in reaching all members of the individual network — including both PC and non-PC users.

In this model, PC communications would be as simple as sending a document to the printer. For example, a user would complete a document and then make a new menu or function key selection in his application to send the document to, say, 10 offices according to a pre-programmed schedule — including five PCs, three FAX machines and two PC/modem/LAN systems. Or two users in different locations would establish a personal E-mail system at a fraction of the cost of a local area network.

The user would send and receive files — be they word processing, graphics, spreadsheet or database files — to and from other PC users within his standard application software; there would be no need to go to a special communications program except to maintain "electronic address book" information on recipients.

And the recipients would not need matching communications software or hardware; a single communications product could transmit data to any telephone-based communications hardware (FAX, modem, voice mail, etc.). The document or file would arrive just as it was sent — formatted within the application, ready for manipulation by the recipient without conversion.

And perhaps most important, the communications system would be an intelligent one, that is, it would offload the work of communications processing from both the user and the PC's CPU to the communications hardware. Communications processing would be executed in the background, while the user and the CPU were free to return to the application file or to any other computer function.

An intelligent communications add-in board would provide a solution consistent with this Intel vision. As both the installed base of PCs and the sophistication of PC users continue to increase, the market for such communications boards also increases.

Until now, the lack of consistent industry standard protocols has been a prime stumbling block to achievement of this vision of intelligent, transparent PC communications.

A new communications specification

Intel joined with Digital Communications Associates Inc. (DCA) to develop a software standard to support transparent PC communications. In August of 1988, Intel and DCA announced the first fruits of that effort: the DCA/Intel Communicating Applications Specification Version 1.0, or DCA/Intel CAS. It is a programming specification in the public domain, allowing developers to turn business applications into communicating applications using a straightforward, standard implementation.

To this project, Intel brings its expertise in PC coprocessing. DCA brings expertise in PC-to-mainframe communications with its IRMA and Crosstalk products. Intel and DCA plan to develop compatible CAS updates that will add interfaces to modems and to DCA's IRMA board, ensuing that PC-to-mainframe communications will be a significant element in the heterogenous communications environment being established under CAS.

These advances will allow users of host-connected PCs to employ a single interface to:

• Communicate directly with mainframes,
   
• Share that information with remote PC users and FAX machines connected only by telephone lines, and
   
• Instantly download communications and continue with other work.

Independent PC software developers endorsing the specification are Alcom Incorporated, Ashton-Tate Corp., Borland International Inc., Conetic Systems Inc., Lotus Development Corp., Microsoft Corp., Novell Inc., cc:Mail Inc., Symantec Corp., WordPerfect Corp. and WordTech Systems Inc. Most of these software developers already have incorporated CAS into their popular PC software.

How CAS works

CAS provides an easy interface for software developers to add communications to existing applications as well as the ability to communicate with non-PC users. E-mail software developers, for instance, are using the specification to extend the power of local area networks (LANs) by providing a bridge between networks, high-speed links to support remote users and a facsimile gateway to people who don't have PCs.

The new programming specification includes core functions of communications scheduling and execution to allow software developers to implement communications the way they currently implement printing — transparently (that is, without the user having to exit the program).

The CAS is implemented via two software components — a Resident Scheduler and Transfer Agents. A software developer using CAS initiates a communication request in his application by calling a Resident Scheduler and indicating the target (distribution list), transmittal document(s) and desired time of communication. The Resident Scheduler can access all available communications technologies, which shields the applications developer<197>as well as the ultimate user — from the details associated with those technologies.

The Resident Scheduler takes information from the application and "looks up" information in the electronic phone book to determine the appropriate communication method and route to each destination.

If the Resident Scheduler is the communications "brains" of the CAS, the Transfer Agent is its "legman". Consisting of a set of code to connect, transmit and disconnect from the recipient, the Transfer Agent ensures that data is transmitted reliably and reports status information to the sender.

The Transfer Agent may execute on the host microprocessor or on intelligent communications hardware, if it is part of the system. Downloading the transfer agents from the microprocessor to intelligent, dedicated communications hardware is what allows communications to take place in the background, while the CPU takes up further tasks. It also allows the architecture to support multiple Transfer Agents and to easily update existing Agents.

Intel was well-positioned to lead the development of DCA/Intel CAS. It followed the same inter-company cooperative process it developed with Lotus and Microsoft to create the Lotus/Intel/Microsoft Expanded Memory Specification. That specification did for expanded memory boards what the DCA/Intel CAS is doing for communications products. It provided the foundation upon which systems vendors could provide compatible products to solve a basic problem in PC use.

In its initial release, CAS supports communications specifically from an intelligent communications coprocessing board — Intel's new Connection CoProcessor board.

The Intel Connection CoProcessor Board

The first intelligent communications hardware product to support DCA/Intel CAS is Intel's Connection CoProcessor board. This communications coprocessing board and accompanying communications software allows its users to exchange files — including binary data, graphics and text — E-mail and FAXes with other Connection CoProcessor board users at a rapid 9600 bits per second. They can also send FAX and E-Mail to all Group III FAX machines and receive FAXes from them.

The Connection CoProcessor board is compatible with IBM PCs, XTs, ATs, 80386 systems and compatibles, and is hardware ready for OS/2. It uses a 10 MHz 80188 microprocessor and 256 KBytes of memory to download communications processing from the main CPU, freeing that microprocessor — and the user — for other tasks.

With this communications coprocessing board, users avoid the complicated protocols and logistics of modem communications and receive feedback on transmission status. Both the transmitting and receiving PCs can be unattended. The product is four to eight times faster than traditional modems in file transfer mode, and that can produce significant savings in long distance telephone charges.

Users can access the Connection CoProcessor board through the simple Intel communications application. That application also serves as the electronic phone book for up to 999 stored distribution lists with up to 999 names in each.

With add-ins and releases for popular software packages including Lotus 1-2-3, SideKick Plus, WordPerfect, and Q&A, Intel Connection CoProcessor board users will be able to send files and FAXes via simple menu or function key commands.

Summary

The Intel Connection CoProcessor is just the first of many hardware products that we will see exploiting the power of DCA/Intel CAS. Already most of the major software developers have shown their support for CAS by actually releasing products that use CAS.

This will have the most impact initially in the facsimile board market. You now have a choice of purchasing many different facsimile boards, each with their own unique interface to software (usually just an ASCII file), or the Intel Connection CoProcessor board which already supports CAS.

Certainly, software developers have signalled they would prefer to see one common interface for all communications products by the way they have so rapidly come forward with support for CAS. During the next six months we believe support for DCA/Intel CAS will increase dramatically.

We believe that eventually most facsimile board manufacturers will be forced to modify their hardware and software to support the DCA/Intel CAS specification. And we see the manufacturers of modems and PC-to-mainframe products being pressured to do the same.

But best of all we see the opportunity for exciting new software applications that will make it easier for us to communicate with each other, in ways not dreamt possible today.