In today’s economy, with the emphasis on safe and successful transactions over the Internet, a reliable and scalable network is key. To become and remain competitive, you must have the ability to move considerable amounts of data back and forth across your network each and every day. Whether you are a small, medium, or large company, Ethernet solutions can provide the technological resources that you need. Ethernet offers functionality and dependability while still serving those increasingly common bandwidth- hungry applications.

The search for the perfect networking solution is a daunting task and may be an unrealistic goal. Networking technologies are changing so rapidly that there always seems to be something better, something faster out there. In many cases, building a completely new network is unfeasible or impractical. If you can restructure a network, the biggest challenge usually lies in maintaining operations without interrupted service. Most companies cannot afford the downtime that restructuring and stabilizing a new network requires. This fact usually means that you have to build on your existing infrastructure. Ethernet solutions on your AS/400-based network present a broad range of technological characteristics at a fraction of the cost.

This Is Ethernet

The Xerox Corporation developed Ethernet in the 1970s, and it was standardized through the cooperation of DEC and Intel. The development team at Xerox Palo Alto interconnected more than 1,000 computers on a 1 km cable using the Carrier Sense Multiple Access with Collision Detection (CSMA/CD) protocol. Ethernet networks transmit information by sending frames of data out to each station on the network. The frame carries an address that targets the destination for the data. After the frame reaches its intended target, the target machine processes the information.

Today, Ethernet is one of the most widely implemented LAN protocols. Ethernet is a flexible protocol offered in a variety of speeds (10 megabits per second [Mbps], 100 Mbps, and 1,000 Mbps) over a variety of physical architectural structures (including Star and Bus technologies). You can run Ethernet over a variety of cabling schemes that include coaxial cable (standard or thin), twisted-pair (shielded or unshielded), or even expensive optical fiber.

IEEE 802.3: More than Letters and Numbers

One of the key elements adding to the popularity and widespread use of Ethernet has to do with the 802.3 standard specification being defined by the Institute of Electrical and Electronics Engineers (IEEE). The IEEE was founded in 1884 to advance scientific theory and innovation. Today, it consists of scientists, engineers, and students who aid in the development of standards for the computer and electronics industry. In particular, it defines the standards for the configuration of Ethernet LANs, the types of media that can be used, and how elements of the network should interact. Here are three important examples of the
802.3 standard:

• Ethernet networks must run at a data rate of 10 Mbps, 100 Mbps, or 1,000 Mbps. Hence, when a station transmits a packet of information onto the Ethernet LAN, it must travel at one of these speeds.

• In a LAN environment, every computer on the network has the ability to send data at the same time, but if two nodes try to send packets at the same time, a collision will result, thus preventing the packets from ever reaching their destinations. Managing and reducing the amount of collisions on a network is a key consideration in Ethernet’s design. Collisions are considered common and are managed in an Ethernet environment by the CSMA/CD protocol. CSMA/CD is designed to effectively manage and restore the network to normal by directing the transmitting computer to retransmit the data until it reaches its intended audience.

• When tracing LAN communications, you need to look at the transmitted frames and understand the information contained therein. The original physical Ethernet frame composition was altered and later standardized by the IEEE.

Analyzing a Frame

Figure 1 displays the format of an Ethernet frame. The starting delimiter field announces the frame and enables all receivers on the network to synchronize themselves to the incoming frame. The destination address is responsible for identifying the recipient of the frame, while the source address identifies the station that transmitted the frame. The byte field consists of two bytes that define the number of bytes contained in the data field. According to the 802.3 standard, the minimum frame size must be 64 bytes in length. Consequently, with the possibility of using two-byte addressing fields, each LLC header and information field must be at least 46 bytes in length to a maximum of 1,500 bytes. This minimum specification ensures that there is enough time to enable Ethernet network interface cards (NICs) for accurate collision detection. Last, the frame check sequence field provides a mechanism for error detection.

This standard Ethernet frame format provides a standardized networking technology, which produces a structured relationship between hardware and software vendors. This relationship is essential to accommodate the growth needed in local and worldwide communications. Standardized networking technologies increase compatibility by defining the physical and operational characteristics of the network each vendor supports.

AS/400 Ethernet Considerations

The AS/400 system supports Ethernet 10Base-T (10 Mbps) and Fast Ethernet 100Base-T (100 Mbps). In planning AS/400 Ethernet support, you must consider a number of factors before choosing the best solution for you. Below are some of the more important considerations for Ethernet on the AS/400:

• Supported IOAs. Before you can create your Ethernet line description, you have to specify the IOA to use with it. Figure 2 shows supported IOAs, speeds, and corresponding duplex modes.

• Frame size. The larger your frame size, the more data your system can pack into it, resulting in increased data throughput. Typically, you set the maximum frame size to the largest size supported by the IOA. Three maximum frame sizes can influence the final selection of the maximum frame size for the connection:

1. The maximum frame size for the Ethernet standard that you select

2. The maximum frame size associated with a source service access point (You can connect SNA to a LAN protocol by using one or more service access points. Multiple Service Advertising Protocols allow you to have multiple connections between SNA and a LAN protocol, thus permitting multiple communication paths between independent applications.)

3. The maximum frame size associated with the controller

During the connection process, your AS/400 uses the smallest value of the three maximum frame values. The final frame size can become even smaller because of negotiations during the connection process or poor bridge considerations:

• Automatic negotiation. The AS/400 supports the automatic negotiation of line speed and duplex mode when used with appropriate IOAs. During automatic negotiation, a device and the switch to which the device connects agree on the duplex mode and the highest common line speed. If the hub does not support automatic negotiation, your AS/400 can detect the line speed. The AS/400 cannot detect duplexing modes, so it will always choose half- duplex.

• Upgrade of existing network. The construction of a brand-new network is something that most network administrators will not do. A more feasible and realistic approach is to build a new network on an existing one because most companies just cannot afford downtime, preferring instead to spend their time upgrading and continuously improving an existing infrastructure. A proven network architecture takes a great deal of money and time to stabilize. In addition, the instant cost of new equipment and educational costs incurred in learning how to use that equipment are other obvious deterrents.

Time to Upgrade?

Is it time to upgrade from 10- to 100-Mbps Ethernet? In many busy networks, it is no longer useful to rely on 10-Mbps Ethernet. With the proliferation of bandwidth-starved applications and multimedia technologies, your networks need greater throughput capacity and performance. There are a number of advantages to upgrading. Most vendors have implemented and tested the new Ethernet standards with their equipment, and prices are steadily dropping. In addition, the variety of vendors from which to choose is growing steadily. The best advice that could be offered is to stay with those that can offer you that complete, end-to-end Ethernet solution.

If you need to upgrade, you should first determine what capacity level your network is currently operating at. As your network’s utilization percentage continues to increase, the necessity and argument for expansion become much more convincing. If network utilization is consistent and you do not foresee the proliferation of larger files and applications, then you could probably remain with the current infrastructure at this time.

However, one of the biggest problems reported with Ethernet technology is an ironic twist on what it provides: speed. Packets are being transmitted so fast with the new

technologies that you may not be aware of collisions or packet loss until it’s too late. The permissible level of failure tolerated by your network administrator and the ability to isolate these problems may not be acceptable. Other common obstacles include the current cabling media, cable-run lengths, and server speed, all of which are affected by or affect the speed of the transmission. To see the topology rules for maximum network distances, see Figure
3.

Switched or Shared?

There has been an emergence of switched LANs and full-duplex technologies in many of today’s LAN environments. They have become a cost-effective way to significantly improve network performance. Originally, LANs began in a rather unstructured fashion, and users were added to the LAN as needed. With Ethernet LANs, each began on coaxial cables, with each station tapping into the cable. As the Ethernet LANs grew, a more structured approach called 10Base-T was employed in organizations. Under 10Base-T, all attaching devices were linked to a repeater. This helped to isolate problems and track inconsistencies within the LAN environment. It also increased the LAN’s ability to participate in a successful expansion. Such an infrastructure is called a shared LAN. Shared LANs have allowed a large number of stations to share a common access method. In smaller networking environments, in the number of both users and attaching devices, these types of environments often serve their purpose. However, when all hosts are connected to the same bus, they are forced to compete with one another for bandwidth.

In contrast, a switched Ethernet LAN replaces the repeater and offers an effective and convenient way to manage dedicated bandwidth in existing Ethernets. Switches are used to increase performance on an organization’s network by separating large networks into many smaller, less congested LAN segments. The bottom line: less competition for resources. Switched networking technology is more expensive, but it is becoming more prevalent today because of its ability to provide efficient management due to less data collisions and network congestion. In addition, switched networking offers better security and the option of dedicated ports to the desktop.

Bringing It All Together

Today’s IT managers are faced with the challenge of streamlining their business processes and enhancing communication to succeed in today’s marketplace. Bandwidth, speed, reliability, and scalability are becoming the all-important buzzwords for our networked world. Ethernet is an immensely popular physical-layer LAN technology because of its ability to offer these characteristics while still pulling your enterprise together in a cost- effective way.