AspeNet Overview

This section provides background on AspeNet and explains concepts you should know to understand how AspeNet works.

Currently, AspeNet Servers communicate with AspeNet Workstations using two different protocols: TCP/IP and MCP/IPX. Both protocols actually consist of a suite of protocols and sub-protocols. Before we get too concerned with such nuances, let’s take a brief look at the evolution of AspeNet network protocols.

AspeNet releases prior to v3.00 made exclusive use of IPX/SPX. Novell, Inc. created this protocol in the early 1980's as a derivative of the XNS (Xerox Network System) protocol suite. IPX is currently in use world-wide, especially in corporate environments.

In 1995, Aspen Research Group released AspeNet version 3.00. This product contained a newly developed protocol, the Message Control Protocol (MCP). Aspen Research designed MCP to be a faster, more flexible alternative to Novell's Sequenced Packet Exchange (SPX) protocol. MCP combined new research in client/server information flow with proven concepts developed by Xerox's Sequenced Packet Protocol (SPP) and the U.S. government-originated Transmission Control Protocol (TCP). MCP replaced the use of SPX in AspeNet version 3.00.

In 1996, beginning with v3.01, AspeNet Servers and Workstations were equipped to communicate using TCP/IP, MCP/IPX, or both simultaneously. The TCP/IP protocol suite, also known as the “internet protocol suite,” was created in the mid-1970s by the Defense Advanced Research Projects Agency (DARPA). By the early 1980s, TCP/IP was available to most universities and major research institutions. It continued to evolve without standard, and, as a result, some of the protocol definitions are over-complicated or obsolete, and therefore often ignored by vendors that develop TCP/IP system software. Fortunately, the basic principles are sound, and a practical set of standards has emerged, allowing widespread compatibility on nearly all computer hardware and operating systems. Recently, the use of TCP/IP has increased dramatically in the world of corporate and personal computing due to the success of the internet.

Either protocol, MCP/IPX or TCP/IP, provides roughly equivalent performance (with a slight edge to MCP/IPX); the decision of which to use is largely an administrative one, driven by issues like cost, availability, administration, and compatibility with existing infrastructure, to name a few.

Transport and Network Protocols

Strictly speaking, TCP and MCP are both known as transport protocols while IP and IPX are both known as network protocols. Because transport protocols by themselves rely on the services of network protocols, you will often see them mentioned in pairs, as in “TCP/IP” or “MCP/IPX.” It is also common to refer to the entire collection of protocols (and there are many we are not going to discuss here) by the network protocol acronym alone. Hence, the use of “IP” or “IPX” is often used as a reference to all protocols, transport and others, that are closely related.

“TCP” stands for Transmission Control Protocol. The main objective of TCP is to create a reliable, well-ordered information exchange between two computers without losing or duplicating data. This information exchange formed by TCP between any two machines (or, more accurately, any two software processes) is known as a “connection” or “virtual circuit.”

“MCP” stands for Message Control Protocol. Its main objective is the same as TCP’s. The primary difference is that MCP controls information flow, sequencing, and delivery verification on a packet-by-packet basis (a “packet” being the basic unit of data transmitted on a network) rather than on a byte-by-byte basis. While MCP treats application data as discreet logical records, TCP treats application data as a stream of bytes having no logical boundaries. Aspen Research Group, Ltd. developed MCP to provide a fast, reliable, and efficient transport service well-suited to the rapid message exchange that typifies computer assisted trading.

TCP and MCP oversee the flow of information between connected computers, taking steps to ensure that all information sent is received in its original form. To do this, TCP relies on the Internet Protocol (IP) layer to send and receive information in discreet units called “TCP datagrams.” By contrast, MCP relies on the Internet Packet Exchange (IPX) protocol layer to send and receive information in “IPX datagrams.” Both IP and IPX are responsible to direct their own kind of datagrams from one machine to another (though neither can guarantee a datagram will not be lost along the way due to network failure or congestion).

IP and IPX view each datagram independently from all others, caring neither for its content nor its relationship to other datagrams. TCP and MCP see a larger picture” of ordered information, but care only about reliably transporting the information from one connection endpoint another. Only the applications which send and receive the information see the “big picture” and know the purpose or meaning of the information.

Protocol Stacks

TCP/IP or MCP/IPX software creates what is known as a protocol stack, a layered interface through which application programs can communicate. The stack” refers to the concept of layering more complex protocols (such as TCP) on top of” less complex protocols (such as IP). The higher-level protocol uses the lower-level protocol to accomplish its objectives. The entire TCP/IP or MCP/IPX stack is, in turn, “stacked” on top of additional, lower-level drivers responsible for managing network hardware.

Network Drivers and Adapters

Hardware-managing software modules are known as drivers. At the request of protocols in the layers above it, a network driver directs network adapter hardware to transmit and receive data to and from a network medium (such as RG-59 coaxial cable, or Category-5 twisted-pair wiring).

Network Types and Frame Types

Physical networks are typically distinguished by the form and electrical characteristics of their media as well as the rules which govern how attached devices share those media. “Ethernet” and “Token-Ring” are examples of some common network types.

Data transmitted across a network medium are encapsulated within structures called frames. Each type of physical network (e.g., Ethernet, Token-Ring) has its own set of one or more frame types. Each frame type provides a means to identify the beginning and end of a data transmission. Most network drivers are able to recognize and use several frame types at the same time.

Packets

The term “packet” has a broad, and sometimes inaccurate, usage. It is commonly used to refer to data undergoing transmission on a network. To draw a more accurate picture, “packet” refers primarily to the data being transmitted, while “frame” refers to the structure of those transmissions.

In terms of organized data, “packet” is often used to describe an IP or IPX datagram. It is also common to use “packet” to describe datagram size because network layer protocols (such as IP and IPX) determine datagram size (as reported by the hardware driver). This relationship perpetuates the general use of the term “packet.”

Layering of Network Architectures

The principle of layering more complex protocols, services, or hardware facilities on top of less complex forms is common in network architectures. The following figure illustrates the logical relationship of the software and hardware layers described during this conceptual overview, giving a representative implementation for each:

Application (AspeNet Server)
Transport Protocol (TCP or MCP)
Network Protocol (IP or IPX)
Network Adapter Driver (NE2000.COM)
Network Adapter & Medium (NE2000 Adapter & Wiring)

When an application originates a message, it is processed by transport protocols and passed down to the network protocol in the form of packets. The network protocol in turn requests the network adapter drivers send appropriate frames of data across the physical network to the intended recipient's machine (or, if the recipient is not connected to the same physical network, to a router machine that forwards the packet to the recipient’s physical network). Once these packets hit the physical network, they make their way to the target machine.

Once received by the target machine, the network adapter and adapter driver sends the message in frames to the appropriate network protocol. The network protocol converts the frames to packets and sends the packets to the appropriate transport protocol. The transport protocol then orders and reassembles the packets into a complete message for delivery to the receiving application.

Usage and Configuration Issues

Forming a connection between two software computers (a workstation and a server) requires proper computer configuration. As you address computer configuration, you should have a good working knowledge of the network protocols and their standards. The following sections provide some general principles about protocol standards and physical networks.

The configuration on the computer that originates a message must be identical to the configuration of the computer that receives the message. That is, a workstation using TCP/IP protocols cannot communicate with a Server that is using only MCP/IPX, and vice versa. However, it is possible to configure a Server with both TCP/IP and MCP/IPX, allowing workstations to communicate with it using either protocol.

At the driver level, any two machines that are connected to the same physical LAN segment must recognize and use the same frame types. Typically, all machines on a network segment are configured to use the frame type of a particular protocol. During configuration, then, one of your first tasks is to associate (or bind”) a computer to the frame type of the protocol you will be using.

The following table identifies industry-standard frame types for IP and IPX protocols on the Ethernet and Token-Ring networks:

	IP	IPX
Ethernet	Ethernet_II	Ethernet_802.2 Ethernet_802.3 Ethernet_II Ethernet_SNAP
Token-Ring	Token-Ring_SNAP	Token-Ring Token-Ring_SNAP

Network Cable

Network cable connects your server(s) and workstations together. All computers (server(s) and workstations) should be turned off before connecting or disconnecting network cabling. Be careful not to crimp network cable, damage its insulation, or run the cable near heat sources, fluorescent light fixtures, heavy industrial machinery, or office machinery. Some of the more common types of network cabling include:

Thin Ethernet

Twisted-Pair Ethernet

10 Base T

Thin Ethernet

Thin-ethernet cable is coaxial cable. At each computer on the network, the coaxial cable plugs into a T-connector. A T-connector has three connections, one of which plugs into a network card in the computer; the others allow you to continue or terminate the network. To continue the network, you connect another segment of coaxial cable to the open plug on the T-connector and run the cable to another computer. To terminate the network, you connect a grounded or a conventional 50 (ohm) terminator into the open plug on the T-connector. Both “ends” of the network must be terminated. One of the terminators must be a grounded terminator. The grounded terminator must be connected to an earth ground.

Twisted Pair and 10 Base T Cabling

Twisted Pair (also known as Twisted Pair Ethernet) and 10 Base T network cabling are essentially the same. Both look like the cable that runs from your telephone to the wall jack. The difference is, these network cables have eight wires, so their jacks (RJ-45 jacks) are larger than their telephone counterparts (RJ-11).

Network Cards

AspeNet works with most network cards as long as they support IPX and TCP/IP. Please read the installation instructions that come with the cards you purchase for information on installing the cards in your computers. Additionally, you should read the instructions on configuring your network cards.