Networks too are highways, with many similar characteristics and tradeoffs. Flexibility, conformance to standards, efficiency, ease of use, security, control, management, cost, and overall performance, to mention a few. Balancing and controlling them end to end will be crucial in building a successful network infrastructure. Will it take you where you want to go? Can it handle new requirements without being torn down and rebuilt? (Dragster engines are torn down and rebuilt after each 1/4 mile of usage, while some engines can go 100 000 miles without a tune-up.)

While extreme, this comparison points out the necessity of looking at all of the factors necessary in choosing the right technology for networks and highway systems alike. Choosing carefully is critical, since as organizations move toward the network computing model, network infrastructures will become your key competitive advantage. You must therefore choose a technology that meets these new challenges and performs seamlessly from the backbone to the desktop. This choice is not always easy, as there are a number of competing technologies available that seem to meet the challenges of these new networks, including Asynchronous Transfer Mode (ATM) and switched Ethernet technologies such as Fast Ethernet and Gigabit Ethernet.

ATM has been accepted by most analysts, vendors, and users as the backbone of choice for future high-performance networks. A recent Gartner Group flash dated May 24, 1996 stated that Gigabit Ethernet should not be considered as a replacement for enterprises looking to implement ATM. The reasoning used has to do with the "divergent networking approaches, using the two different philosophies of high-bandwidth vs. managed bandwidth." It goes on to add that the "performance of Gigabit Ethernet will follow that of 10 Mbps and 100 Mbps Ethernet, with a maximum sustained rate of 40 percent utilization." It concludes that "ATM will remain a superior technical solution for LAN and WAN backbones in both new and in-place networks (0.8 probability)."

Acknowledging the strong consensus supporting ATM as the backbone of choice, this paper contrasts two of the several technologies available for desktop users today and in the future. 25-Mbps ATM comes out on top over 100-Mbps Fast Ethernet for most customers' current and future desktop application requirements. Meeting these needs gets down to balancing all of the components in the end-to-end solution that will deliver the applications. Delivering high-quality applications is how networks will be judged. Does it handle all the delivery requirements of current and future users? Does the architecture allow for multimedia applications? Can voice, full-motion video, imaging, as well as high-priority data applications be accommodated on the same network effectively? Has the network been architected from the "frame" out to support these applications, and is this frame based on industry standards?

It's not all about raw speed. Control, flexibility, adaptability, cost, and manageability are all significant aspects. Balance is an essential aspect in the overall performance of any networked system. Matching the technology with the user's requirements and applications will be the challenge. ATM's scalability coupled with IBM's Switched Virtual Networking will allow you to match the application needs of all of your end users with the best-fit technology across a common architecture.

In this paper we will show you how 25-Mbps ATM to the desktop gives you the best balance and support for current and future applications. It has the robust architecture, management, ease of use and, yes, performance to accommodate the time-sensitive and data-intensive applications of the future. It also has the best scalability, meaning it offers you the most flexibility to adapt to changing conditions in the same standard technology.

Finally, ATM is the only technology that spans the local area, campus, and wide area networks and crosses international boundaries as an end-to-end, standards-based solution. In a networked computing world, this reduces your risk on issues of interoperability, availability of choices, network skills, and long-term viability. We believe you will agree that a desktop ATM decision now will be the right choice well into the next century.

Sage Research of Natick, Massachusetts, finds in a recent study that most demand for high-speed LAN adoption is driven by the use of multimedia applications. Based on a survey of 175 network professionals, the report found 72% of respondents agree that their network goal is to support delay-sensitive traffic, such as voice or video, over their LANs by 1999. Today's traditional networks cannot support these delay-sensitive applications and thus require a faster, more reliable alternative.

How fast do these networks need to be?

Figure 1 shows a projected implementation of voice, video, and data technology. In this scenario, the user is transferring a data file, maintaining a two-way videoconference with a fellow employee 400 miles away, talking on his computer-based telephone to a client on another continent, and running two separate sets of video footage of a recent news event that might affect his business. To do much more than this at the same time, the user needs more than technology. He needs super-human concentration.

Figure 1. 25-Mbps ATM to the Desktop, 1996 to 2000 and Beyond, Average Users

As you can see, even with all of these applications running simultaneously, the required bandwidth is still within the capability of 25 Mbps.

But speed is not the only concern. In an office that depends increasingly on these applications, reliability is just as important. ATM technology is the only reliable solution to support video, multimedia, and other time-sensitive applications.

What about Fast Ethernet?

The Sage report states that Fast Ethernet vendors "must position their product as a less expensive alternative to ATM while either responding to the video challenge or targeting other applications."

ATM has met the video challenge. The pages that follow examine the application capability of 25-Mbps ATM to the desktop, as it relates to actual user implementations.

Then it must be expensive, right?

Not so. As we show in the Pricing section of this paper, the price of a 25-Mbps ATM solution is less than or equal to currently available Fast Ethernet solutions.

Health and Insurance
One of the largest ATM networks in the world is installed at Kela (Helsinki), Finland's social security insurance organization. They have recently completed the initial installation, which links 11 locations through Telecom Finland's fiber network and IBM 8260 ATM switches. The initial network consists of over 1500 workstations with 25-Mbps ATM desktop connections. In the end, over 5000 desktops will be linked through Kela's integrated LAN/WAN ATM network.

Image-processing applications are the motivation for building this infrastructure. Kela decided that keeping insurance records in hard copy was too cumbersome and expensive, as employees had to wade through massive filing cabinets to retrieve information. Kalle Helameri of Kela's system support group had a vision that service representatives would be much more productive if they could retrieve documents using image-processing applications through a network. Service representatives can then find the information they need, independent of its location. In this distributed organization, this is a major benefit.

Customers of Kela benefit from this reengineering, too. With image files instead of paper, customers can go to any Kela office for help. Quality of service will be improved as well, since the most knowledgeable person can access the information and deal with the questions, as opposed to being tied to the physical location of information. Kela is also working on distributing workloads from busy offices to those with excess capacity. In the past, customers have been assigned to a particular office where their records are stored. When that office was too busy, customers have had to wait.

LAN Emulation, a technique that allows any application to use an ATM network, is being used successfully at Kela. It allows normal source-route bridging techniques to be used to the hosts. They currently use four emulated LANs across the ATM network. Overall reliability has improved consistently and is above 99.9%.

Helameri is delighted with 25-Mbps ATM performance. He does not foresee any application that would benefit at this point from more bandwidth to the desktop. He is confident that they have enough bandwidth for the future and is planning to add an additional three thousand 25-Mbps ATM desktops.

Manufacturing
The manufacturing sector is rich with application solutions for which ATM is the best technology. As a result, many manufacturing companies are upgrading their infrastructures to enable and exploit ATM.

Key early applications include Computer Aided Design (CAD), Computer Aided Manufacturing (CAM), training, conferencing, and manufacturing automation. Many manufacturers require the ability to simulate their designs for safety and performance. Such requirements often involve both large amounts of data as well as image processing. In many cases, high-performance or parallel-processing computing capacity is required and needs to be accessible to engineers and product designers located in various parts of a large campus or wide area network.

As with other industries, the potential for network consolidation of voice, video, images, and data traffic types over ATM is a key factor in deployment of ATM infrastructures. ATM is the seamless transport from the desktop across the campus into the wide area, which allows for savings in network consolidations and simplified network management.

Chrysler Corporation's engineering Ethernet network was running out of gas. They evaluated the alternatives including switched Ethernet and Fast Ethernet. They concluded that these solutions would not meet their application requirements or consolidation plans for the following reasons:

• Ethernet and Fast Ethernet do not support the Quality of Service required for multimedia applications including voice and full-motion video. The large data files that are normal in CAD applications were choking their network.
• Ethernet could not span the wide area components of Chrysler's network as well or as seamlessly as ATM.
• When they evaluated cost, desktop ATM compared favorably with Fast Ethernet.

As a result, Chrysler is embarking on a very large 25-Mbps desktop ATM solution. Several thousand desktops will be connected to ATM switches in production mode by the end of 1996.

This network will support Chrysler's engineering group with the performance necessary to support several important applications, including CAD and CAE (Computer Aided Engineering), a data-intensive application used by engineers. The network will also support full-motion video. Video applications include viewing the Chrysler Employees Network, a TV-type, news-oriented broadcast to employees at their desktops. Videoconferencing will be used for collaboration and as a meeting facilitator. Other video applications include employee training and CAE crash and wind-flow simulations.

The results to date have been excellent. "The most obvious benefit is that things are working over ATM that did not work over Ethernet," says Bob Rodgers, network designer for Chrysler engineering. "We've even found that in certain conditions, users can mount a server over ATM and run processes faster remotely than they could on their workstations."

With ATM, the Solutions Are Simplified
Chrysler's Bob Rodgers has shown that ATM has simplified the design and overall management of the network. Running large or small networks with multiple application requirements normally leads to layers of complexity. Rogers' desire to simplify this complexity contributed to his choice of ATM. The goal was an end-to-end, ATM-based communications infrastructure.

Education
Educational institutions and other industries with training needs represent a fast-growing area of ATM applications. Many applications have been implemented in a variety of disciplines including:

• Engineering and science research—highly parallel supercomputing
• CAD/CAM
• Internet access, digital electronic libraries
• Computer-based training and self-study
• Videoconferencing
• Video distribution—just-in-time training

At Cornell, IBM is focusing on several aspects of ATM campus infrastructure in a partnership covering:

• Ethernet-to-ATM LAN migration
• Campus telephony services (voice over ATM)
• Videoconferencing both on campus and between campuses
• Distributed supercomputer networking

Government
In government, ATM applications are also being built around multimedia. Governments are looking at multimedia enhancements to many public service applications as a means to improve their service levels while maintaining or reducing staffing costs.

Examples of such applications include kiosks for tourism and government services such as license application and renewal. Other institutions are creating Internet web pages to provide information about their agencies, cities, and services.

At the state and federal level, many government groups are sponsoring other less government-oriented programs including grants for solution development.

Two of these are unique for 1996:

• The National Information Infrastructure initiative has launched Internet 1996 Work Exposition, a high-bandwidth assemblage of events, global networking infrastructure, and Internet pavilions.
• IBM support for the 1996 Olympic Games in Atlanta involved deployment of ATM infrastructure and some ATM applications.

Although not all solutions involved will be ATM-based, key, time-sensitive applications including conferencing and kiosks have been ATM-enabled.

Banking, Finance, and Security
As the banking industry becomes increasingly competitive and takes on a global view of business, pressure to reduce costs through consolidation of network and IS facilities, through deployment of "virtual branch expertise," leads to more networked applications and increased deployment of improved user interfaces and training applications. The need to get an answer quickly results in such applications solutions as conferencing and branch transaction networks by which small local branches can offer full services to their customers without having all staff functions at each location. In some banks, kiosks are being implemented to permit customers access to their financial information and then a conference with a loan officer or financial advisor at a central location. Other banks are partnering with public network operators to provide online banking services including investment services with applications that involve digital libraries and conferencing in addition to traditional text.

Large and small banks alike are concerned about providing critical site backup. For a bank with the need for major online and batch processing to be reconciled on a daily basis, little time is available for backups unless the process can be automated and sufficient bandwidth is available to transmit essential records. ATM's scalable bandwidth in both the LAN and WAN make it the ideal choice for solving these time-sensitive, critical backup needs.

While most of us are familiar with local retail banking, and have come to expect fast, seamless transactions, a significant set of applications is required to support the investment and money management applications of a bank. One example of such back-office applications is the trader's workstation that permits a trader to be aware of events and stock market events through online news services (for example, desktop news with access to networks such as Reuters), to be able to engage in conferences with industry and financial consultants via conferencing, and to be able to monitor the stock exchange trading floor and submit orders. A typical Wall Street trader's workbench might look like the configuration shown in Figure 2.

Health Care
Health care networked applications are amongst the most exciting because of the potential of using technology to help people. Radiological imaging, including X-rays and MRI images, require large amounts of bandwidth to transport. ATM is being used as a transport to move these images around hospital buildings, campuses, and a metropolitan area network, and eventually across the global wide area.

To the desktop, 25-Mbps ATM can support these requirements with 155-Mbps connections to supercomputers where wedges of scans can be assembled into complete images.

Video-based remote consulting and collaboration, neurosurgery visualization, picture archiving and communications systems (PACS), and virtual brain surgery and education are some of the other potential visually oriented, time-sensitive applications in various planning and implementation stages now.

Duke University, Dayton Area Hospital Network, USC Medical Center, and the University of Pennsylvania's Children's hospital are just a few organizations who have begun work in the above application areas.

Additional Opportunities
There are numerous applications in other industries that have installed ATM for new and future applications as well as for establishing the infrastructure for growth.

The infrastructure they are building today will accommodate the applications we are currently using, the ones we plan on doing, and the ones we haven't yet thought of doing. They are on the way. The future is full of exciting multimedia applications.

IBM has received numerous requests from its customers for First Virtual's high-quality desktop video networking solutions for applications like video conferencing, collaboration and distance learning.

Using First Virtual's family of products and IBM's broad range of ATM desktop and enterprise offerings, IBM customers will be able to bring high quality video and sound to desktop computers in real time. This real-time interactivity allows customers to improve cycle time to market for new products, rapidly disseminate critical information, and significantly reduce travel time away from the office.

First Virtual also announced it has worked with IBM to complete full interoperability testing between its family of desktop video products and IBM's ATM products, as well as video conferencing products from PictureTel. The testing was conducted at IBM's labs located in La Gaude, France in addition to First Virtual's headquarters' labs in Santa Clara, California. The testing was completed successfully since all the companies fully adhere to the ATM Forum standards.

According to Mark Knittel, director of product marketing, IBM Networking Hardware Division, "High-quality video conferencing and video collaboration products from First Virtual, combined with IBM's scalable desktop and backbone solutions, fully exploit ATM's asynchronous capabilities. IBM and our customers are excited about the productivity improvements that merging of voice, video and data bring to the desktop. IBM's ATM products provide the enterprise infrastructure required for maximum customer benefit from these applications."

A True End-to-End Solution
From the desktop to the workgroup across the campus and the wide area, a common communications management and transport protocol is the standard. This LAN-to-WAN integration simplifies the design, modification, and management of the entire network. It just makes it easier to drive.

The Challenges of Today's Shared-Media Environment
The networks that have grown and changed to meet changing demands in our organizations have also become complex and costly to manage and maintain. Increasing numbers of users, new bandwidth-intensive applications, increasing Internet access through the LANs in our offices, and more workstations attaching to each network segment have all contributed to less available bandwidth per station and inadequate performance on some networks.

Dividing large networks into smaller ones joined by bridges or routers and increasing the power and number of servers are only temporary fixes. Worse, all of these measures tend to add complexity to networks by the introduction of dissimilar technologies, therefore increasing the cost of network operation. In all cases, the ongoing cost of operating and managing a network greatly exceeds the cost of the hardware. Reducing these costs is a major consideration as we make decisions about the technology we choose for our networks.

For the workgroup, the switched Ethernet and switched Token-Ring implementations coming to the market are very effective in running legacy applications. The addition of an ATM uplink to an ATM backbone helps relieve the access bottlenecks from workgroups to servers and other resources located on the backbone. Not all network users will require the optimum solution: some legacy applications, such as data entry, will always be easily accommodated on today's low-cost, shared-media LANs. But for the knowledge worker of the future, whose workstation will serve as telephone, storage facility, blackboard, fax, and virtual conference room, only dedicated-media ATM has characteristics that will make it the optimum solution for these new applications. Indeed, the sooner you migrate to ATM, the greater the benefits in terms of reduced complexity and lower cost of operations as well as superior performance.

Initially, these LAN switches were very expensive and were used primarily to meet the bandwidth needs of high-powered workstations and demanding graphics applications. More recently, however, they have become an affordable, alternative method of relieving bandwidth congestion in large workgroup networks running legacy applications. They are simple to install in that they do not require new adapters or protocol stacks, and they employ existing building wiring for connection to the switches. They can also reduce network complexity by reducing the number of bridges and router connections in the network, so that there are fewer entities to configure, monitor, and manage.

An attractive aspect of these two, similar technologies is that they provide solutions for congestion problems on existing LANs. For example, a shared-media LAN of 60 users that is suffering from client/server traffic problems can be divided into 10, six-station, shared-media LANs that are attached to ports on a LAN switch. The servers are isolated on ports of their own and their adapters are reconfigured to operate in full-duplex mode. This configuration reduces congestion significantly while preserving the investment in adapters, hubs, and building cabling. As the traffic continues to increase, each workstation can be attached to its own port on the switch to maximize the bandwidth capability of each device. Finally, high-use resources such as servers attached to their own ports can be operated in full-duplex mode, effectively doubling the available bandwidth at that port. Such a change improves server throughput and limits the number of changes needed in the network to bring performance back to an acceptable level.

Perhaps some of the other issues of Fast Ethernet are even more significant. Its lack of scalability is a concern in networks where growth is expected. It is also less robust because its inability to provide multiple uplinks seriously limits its fault tolerance, redundancy and performance. There is also the issue of distances that must be understood. The distance limitations of Fast Ethernet significantly hinder its ability to be a backbone technology. While this may not be a problem in all installations, the total, single collision-domain length is reduced to 250 meters (820 feet) from 2500 meters (8202 feet) for 10BASE-T Ethernet. This is a function of the reflection timing necessary for CSMA/CD technology. Related to this, the acceptable number of repeater hops has been reduced from 4 to 2.

Time-sensitive applications do not move smoothly across the network because there is no congestion control before establishing the end-to-end connection, and data might be lost. Whether in the backbone or desktop LAN environment, this frame-based technology does not have the isochronous capabilities required by the emerging multimedia applications. This can be best understood by reviewing the following points.

Quality of Service
Quality of service (QoS) is an important issue, especially in a multimedia environment. The degree of network service, including network delay, jitter, and cell loss ratio, are very important when dealing with voice, video, and other time-sensitive applications. Fast Ethernet has no notion of QoS and therefore compares to an Unspecified Bit Rate (UBR) in ATM, ATM's lowest QoS. ATM has three higher service levels defined in its architecture, as described in "ATM Quality of Service Parameters".

Frames versus Cells (Ability to Multiplex)
With Fast Ethernet, frames sizes will still vary between 64 and 1516 bytes. Many applications use the maximum frame size available and there is no mechanism for preventing a maximum-size frame from coming into a network. When maximum Ethernet frames enter a network, they can cause slowdowns for time-sensitive traffic such as multimedia applications. In this example, if time-sensitive traffic were behind a large frame on Ethernet, it would incur the delay for processing the large frame. It would be similar to a small sports car being slowed down behind an 18-wheel tractor trailer, but worse.

In an ATM network, cells are all of a fixed length at 53 bytes. As shown in Figure 3, a 1516-byte frame is 28.6 times larger than an ATM cell.

Figure 3. ATM Cell Size versus Ethernet Frame Size

ATM cell size was chosen with multimedia support in mind. It is significantly more capable of prioritizing traffic on a cell basis. ATM can multiplex and prioritize over 28 separate traffic cells while one Ethernet frame would be handled in a serial fashion. ATM's ability to prioritize traffic is therefore greater because of this difference in granularity.

Ethernet technology is more serialized by design. It will therefore tolerate delays. In addition, any delays incurred by large frames are repeated at every node in the network.

Wiring Differences
25-Mbps ATM can be supported with a single pair of category 3 unshielded twisted pair (UTP) wiring, same as 10BASE-T. However, Ethernet, 100BASE-TX, requires a single pair of category 5 UTP. While most users are installing category 5 wiring today, it is estimated that over 30% of installed wiring is category 3, which cannot support Fast Ethernet. Because of this limitation, another standard is emerging called 100BASE-T4, which uses two pairs of category 3 wiring. However, few vendors support this new standard.

As you probably realize, installing new wiring and dealing with network delays due to inferior technologies can be very expensive to your business. So what's the solution?

A Connection-Oriented Protocol
Traditional shared-media LANs use a connectionless protocol that has proven adequate for most data-oriented applications. Even though switched Ethernet and Token Ring are dedicated-media, switched solutions, they are still fundamentally connectionless in their operation because they are based upon the original, shared-media protocols. Currently, only ATM provides the connection-oriented environment required for the emerging multimedia applications. In addition, ATM's environment offers considerable benefits for running connectionless legacy applications.

A network is a traffic-control system that manages the delivery of goods to and from devices attached to the network. Like the traffic-control system of a city that defines the rules for the delivery of goods across its infrastructure of streets and highways, each network protocol has its own set of rules.

Connectionless Protocols
Let us assume for a moment that the traffic system of your city is based on Ethernet's (or some other connectionless protocol's) set of rules. Each driver starts for a destination as soon as the street appears to be clear of traffic. The driver has no knowledge of the route or of others who might also want to use the route at the same time. It is a hit-and-miss system with congested traffic, collisions, and restarts, with no guarantee of arrival times.

Connection-Oriented Protocols
Let us now assume that our traffic system is based upon ATM's set of rules. Each driver calls ahead and requests a route before the journey begins. The driver gets a reserved route for the trip, along with a lane on the highway wide enough for his or her vehicle. Both the lane and its width are reserved exclusively for the duration of the trip. Nothing has been changed in the size of the highway (cabling infrastructure). The ATM traffic system automatically picks the most expeditious route through the maze of streets and highways and continuously reconfigures the route as traffic patterns change.

In a connection-oriented environment, data is kept in the end-station storage media until the connection to the receiving station is made. Therefore the network is not burdened with the management of data that is en route, thus allowing most efficient operation that is simpler, with predictable destination arrival times. This is why ATM has caused so much excitement in our industry.

Because all legacy LAN applications have been written for a connectionless environment, it is necessary to map connectionless to connection-oriented sessions to use the applications over ATM. Two mapping techniques—Classical IP and LAN Emulation—have been defined by industry standards to facilitate interoperability among various vendors' products. The migration path afforded by LAN Emulation might actually slow down the development of native ATM applications because the benefits that ATM brings to the legacy applications are substantial. Simply by moving the applications into ATM's connection-oriented environment, the classical LAN applications run better because they can take advantage of ATM's higher and dedicated bandwidth. In addition to improved performance, inherent ATM characteristics such as the ability to employ virtual LANs reduce the cost of operating and managing legacy applications when they are run over ATM networks using LAN Emulation.

Speed
In LAN switching, each frame has a different length and destination. The processor in the switch must make an individual decision for every frame. Therefore, the actual throughput capacity of a switch is directly linked to its processor's power and limitations. Techniques like cut-through switching, where transmission is started as soon as enough bytes have been read to recognize the destination address, can improve the latency of the switch on a port-to-port basis. However, filtering, port-speed adaptation (10 Mbps to 100 Mbps, for example), and high error rates on the media often prevent cut-through switching.

In ATM, the data is split into fixed-length cells of 53 bytes each, where a header of 5 bytes contains the routing information. The characteristics of the connection are negotiated ahead of time and, if the network can guarantee the quality of service, the call is accepted and the path is established. Then, the cells are transmitted at hardware speed without the need to reexamine the contents of the cell or perform intermediate store-and-forward actions between the source and the destination.

A Multiplexing System
In a LAN environment (either shared or switched), applications in a workstation or on a server take turns sending data onto the media. Sometimes a low-priority file transfer might delay the transmission of a short frame that requires limited delay. This delay will be repeated at every network node and affect the performance of the network.

In ATM, because elements of information are split into 53-byte cells, cells from different sources can be interspersed and queued according to their individual priority. Thus, fixed delays can be respected, and quality of service can be set according to the application's requirements rather than those of the adapter.

Superior Bandwidth Capability
ATM is, by its architecture, a full-duplex, switched solution. Although some Ethernet and Token-Ring switches and adapters do have full-duplex capability, the LAN switch must accommodate diverse attachment port characteristics and will act as a store-and-forward gateway between ports. This reduces the real capacity and bandwidth of the network.

In ATM, bandwidth is a parameter in the definition of a switched virtual circuit and is independent of the physical attachment: there is no need for intermediate buffering. In fact, if a physical link reaches capacity, additional connections can be added to expand the bandwidth and support additional traffic. Because this capability is one of the fundamental building blocks for high-quality videoconferencing, ATM networks not only provide better throughput for legacy applications, they also provide the infrastructure for emerging applications.

Backbone Access
Most Ethernet and Token-Ring switches are essentially multiport bridges. They cannot use multiple uplinks, and the aggregate switch capacity must remain commensurate with that of the uplinks.

Because ATM is a connection-oriented protocol, bottlenecks between the workgroup switches and the backbone are easily removed by installing additional uplinks between the workgroup switches and the higher speed backbone. ATM switches are able to set virtual circuits over diverse routes according to the current network capacity usage or according to the availability of a specific path. This not only increases the possible link bandwidth but also offers the possibility of bypassing a failing element. As the number of users per floor increases in an end-to-end ATM network, the bandwidth per user need not be affected because of any limitation on uplink bandwidth. Furthermore, installing additional uplinks is simple and should cause little or no disruption in the network. The ability to provide multiple links guarantees uninterrupted service to end users.

Quality of Service
ATM, with its multiplexing architecture, is designed to support traffic with various bandwidth, jitter, and delay requirements. This design feature allows ATM networks to support voice, video, and data multiplexed on the same links. Quality of service is established at the time that the connection is made. Implementing quality of service is dependent upon ATM being a connection-oriented protocol. The ATM Forum has defined four quality-of-service types that are architected to handle the different types of traffic.

Constant Bit Rate (CBR) and Variable Bit Rate (VBR) are particularly well suited for supporting applications with stringent requirements for quality of service, such as multimedia transmission or high-quality videoconferencing.

Multicast Capability
In networks of LAN switches, filters can be used in the switches to control broadcast traffic, but they have an adverse effect on the overall performance of the network.

Multicasting capability and LAN Emulation, which builds broadcast trees for Virtual LANs (VLANs), are foundations of both video distribution and videoconferencing and are exclusive features of the ATM architecture. Unlike the recipients of a broadcast message in a shared-media LAN, only those who want the message will receive it. Because traffic is connection-oriented, no network resources are wasted, and there is no danger of a broadcast storm. In VLAN implementations over ATM, multicasting and LAN Emulation define precisely which stations should receive the broadcast data. In addition, the broadcast manager of LAN Emulation can be augmented with filtering capabilities to reduce the amount of overhead data generated by chatty LAN protocols such as AppleTalk.

Low Latency
The emerging bandwidth-intensive, isochronous applications can work only in an environment where the latency of any one switch is predictable, constant, and extremely low as opposed to variable and unpredictable. In environments where variable-length data and per-frame filtering are employed, latency is adversely affected. In networks where ATM is employed end-to-end, the transit time between any two points on the network will always be the same, so the response time in a large network will be predictable and constant.

Reducing Network Complexity
It is generally estimated that up to 70% of the cost of network ownership is in the cost of operating the network. Therefore, the simpler the network is, the less costly it is likely to be to operate. If we look at a typical LAN environment today, shared-media LANs are joined to backbones, which are often running a different LAN protocol, by bridges or routers. Connection to the WAN is generally through routers as well. Bridges and routers are high-maintenance items, especially in networks with many moves, additions, and changes. Configurations have to be updated, and the network has to be tuned for best performance.

ATM's quality of service and scalable bandwidth virtually eliminate the need for network tuning. Bridges and routers are replaced by simple connections between switches. The result is a network that is more reliable, ready for emerging multimedia applications, and that operates at a lower cost.

Virtual LANs
In current LAN environments, workstations are tied to a port on a specific device so that the functions available to that device correspond to what the network administrator has predefined in the physical network for security or for access to resources. If the user relocates, the network administrator must assign to the new physical port the characteristics that match the user's need. Because affinity groupings are often used as a way of managing networks, when workgroups are reorganized, users have to be reassigned to different physical ports. In traditional LANs, if all or part of the affinity group moves to a different building, the network administrator might have to make physical modifications to the backbone devices (filtering tables or interbuilding links) to preserve the previous capabilities.

VLANs, as implemented in ATM, allow users to belong to several VLAN affinity groupings and share common services no matter where they are physically located in the network. In the VLAN environment, who you work with becomes more important than where you work. When end stations are using LAN Emulation, assignment to VLANs, also called emulated LANs (ELANs), is automatic and is provided by a LAN Emulation Configuration Server. LAN Emulation guarantees assignments to the same VLANs, regardless of the user's physical location. VLANs, because they do not require the intervention of the network administrator or the assistance of a technician to enable and assign a LAN port, can be a major source of cost savings in an environment with frequent moves.

Network Access Control
Since the inception of LANs, designers and network administrators have struggled to find methods of restricting access to only authorized users. Shared-media LANs use intelligent hubs to check MAC addresses against the list of authorized users. When a violation occurs, an alarm is sent and the port where the violation occurred is shut down. This protects the port, but does not really control access to the LAN. Because Token-Ring and Ethernet switches have been designed for performance, they do not perform address checking very efficiently. Most implementations use some form of MAC address frame filtering, which has to be performed on every frame in these connectionless protocols. Obviously, performance can be significantly degraded.

On the other hand, the inherent characteristics of ATM make protecting the network from unauthorized users straightforward. ATM's connection-oriented protocol requires a call to be processed before any connection is established. It is then a simple matter of implementation to check the connection request against an authorization record. This capability can be easily extended to legacy applications because the LAN Emulation server will establish the connection for the application. If the registration is rejected, an alarm is sent and the station is not permitted to use the network. The port is not shut off, and the network performance is not degraded. In addition, ATM allows you to implement other security measures, if needed.

LAN Emulation Benefits
LAN Emulation protocols allow ATM networks to provide the appearance of Ethernet and Token-Ring LANs. While it does not exploit all of ATM's benefits, it is very powerful in migrating to ATM technology and lowering network management costs. High-speed ATM links can be utilized with your current software and hardware, improving performance while protecting your investment. Software investments are protected because application interfaces are unchanged.

Members of an emulated LAN (ELAN) or VLAN need not be in the same physical location, an important benefit in today's increasingly distributed, global markets. In addition, individual workstations may be members of multiple ELANs, providing even more data sharing.

8285 Integrated LAN Emulation Services
The LAN Emulation Server (LES) function of the 8285 gives your existing network transparent access to the ATM network.

Essentially, the LAN Emulation function resolves current Ethernet or Token-Ring MAC addresses by translating them into ATM addresses. It also performs broadcast and multicast LAN services. The system lets you create up to two Ethernet or Token-Ring LANs and concurrently supports both UNI 3.0 and 3.1 LAN Emulation clients (LECs).

The Broadcast and Unknown Server (BUS), also integrated with the LES, has two basic functions:

• It distributes multicast frames to all the LAN Emulation clients (LECs) representing the legacy workstations in the Emulated LAN.
• It forwards unicast frames to the appropriate destination. A LEC sends unicast frames to the BUS if it does not have a direct connection to the LEC representing the destination.

To aid in the configuration of the ELANs, a LAN Emulation Configuration Server (LECS) can be used as a productivity tool. It helps configure LECs by providing them the ATM address of the LAN Emulation Server.

Figure 4 summarizes the functions of IBM's Forum-compliant LAN Emulation.

Figure 4. IBM's Forum-Compliant LAN Emulationb

The integration of the LES and BUS functionality into the IBM 8285 avoids any need for external workstations, which not only simplifies the mode of operation but also reduces the cost of ownership, improves reliability, and eases its configuration.

Network management of ELANs is simplified by increased flexibility in handling moves, adds, and changes. As long as ELAN memberships are retained, no reconfiguration is needed when stations move to new physical locations. Similarly, no wiring modifications are needed to move stations from one ELAN to another.

IBM's Switched Virtual Networking (SVN) strategy uses LAN Emulation but includes additional functions to aid in the migration to ATM networks. It includes Multiprotocol Switched Services (MSS), which adds high-speed switching functions to replace traditional routing services. These functions are aimed at simplifying the coexistence of current legacy LANs with high-speed, switched ATM networks.

Dataquest Forecasts 25-Mbps ATM Market to Grow Twelve-Fold in 1996
A recent Dataquest study showed that the port rate of 25-Mbps ATM in the fourth quarter of 1995 increased 109%. And 1996 predictions for growth are expected to grow at 1109%! Actual shipments in 1995 were estimated at more than 11 000 units with a projected growth to over 550 000 units by end of year 1998. As a result, many newer entrants in the field, including Fore Systems, Cisco Systems, and others are giving their support to this technology and allowing more choices. According to Dataquest, IBM is the number 1 vendor in this opportunity to date with a 45% position.

Much of the strength of this technology comes from support and collaboration across the computing industry. This cooperation was first demonstrated in an organization called the Desktop ATM25 Alliance, in which over 50 organizations participated. Their mission was accomplished in 1995 when the 25-Mbps ATM standard was adopted by the ATM Forum. While IBM is proud of the contributions and work it performed as a member of the Alliance, it was the work of the entire group that made it successful. Below are some of the companies that participated in the Desktop ATM25 Alliance.

As a user, you can benefit from true open standards technology, which will provide you with competitive choices and pricing in various segments of desktop ATM.

Figure 5. Average Price per Port on Switch (Adapter Extra) as of 2nd Quarter 1996

IBM has packaged a special price of $495 "per seat," which includes a 25-Mbps, full-duplex adapter for ISA or PCI buses and a switch port on the IBM 8285 Workgroup Switch, which includes a Forum-compliant LAN Emulation Server function.

Shared Fast Ethernet pricing is not used in this comparison, since it is not a dedicated switch port and cannot support time-sensitive applications.

Token-Ring and Ethernet technologies will continue to have long lives in environments that are not time-sensitive or data-intense. ATM will add balance to existing networks taking on the additional challenges of new applications that require more capability. These new applications are the difference makers that will give your organization a solid competitive advantage.

It's like building a highway system to and throughout your organization. Can it be managed? Will it perform adequately? Will it scale? Will it support the vision of the changing organization? Will it last?

Our advice: Choose a solid foundation. Choose one built from the ground up to support the new, time-sensitive multimedia applications. There are some ideas of taking "ATM-like" functions and adding them to older LAN technologies. These "add-ons" will attempt to do in the layer 3 network layer, what ATM already does in hardware. These are purely add-ons to 20-year-old technologies. They are not integrated by design, and lack end-to-end standardization.

Why are these add-ons being considered? Applications are the key! The time-sensitive multimedia applications are real today. They are demanding new capabilities in the networking infrastructures we are building. Should we take a Model T and add on a supercharger? If we do, won't it still be a Model T?

For comments about this white paper, contact:

Comparison Table (Summary)

Figure 6. Comparison of High-Speed Networking Features

Strengths and Weaknesses of New Switched Desktop Strategies

ATM Quality of Service Parameters

Variable Bit Rate
Like CBR, VBR is a reserved bandwidth service. The network allocates resources to the end station at call setup in response to the traffic parameters requested by the end station. However, in the case of VBR, in addition to a peak rate, a sustainable rate and a maximum burst size are established. The sustainable rate is the upper limit of the average rate, and the maximum burst rate limits the duration of cell transmission at peak rate. These additional parameters allow the network to achieve statistical multiplexing by allocating fewer resources for the connection than would be required by the peak cell rate.

In most campus environments today, the majority of traffic is data transfer that, for the foreseeable future, will operate over ATM using either LAN Emulation or Classical IP mode. These legacy applications are not able to specify the quality of service that they will require. The ATM Forum is proposing that this traffic employ either Unspecified Bit Rate (UBR) or Available Bit Rate (ABR).

Unspecified Bit Rate
UBR is a non-reserved bandwidth service. The cell loss ratio is unspecified, which means that the network is not required to provide resources for a proposed UBR connection. No flow control parameters are specified in the ATM Forum for UBR service. Consequently, when UBR service is employed, cell discard seriously impacts the overall performance of the system. For example, a single cell discarded in a 192-cell packet (the default size for an IP packet when using Classic IP over ATM) triggers retransmission of the whole packet. The network has transmitted 191 cells needlessly. To avoid wasting network resources in this way, early packet discard and partial packet discard can be implemented in any intermediate node (switch) of the network. If a switch recognizes that a cell has been lost, it discards the rest of the packet. If a sending station fails to acknowledge a congested condition, the incoming switch in the network will reject the packets until the congestion disappears. When early packet discard and partial packet discard are implemented in conjunction with virtual circuits, fairness and hop-by-hop backpressure mechanisms insure loss-free UBR operation.

Available Bit Rate
ABR service can be seen as a mix of reserved and non-reserved bandwidth service. Periodically, a connection polls the network and, based upon the feedback it receives, adjusts its transmission rate. Polling is done by Resource Management cells sent by the source and looped back at the destination so that the network elements and the destination can provide feedback information. In addition, network elements can create and insert RM cells in the backward direction to provide feedback to the source more quickly.

Feedback can be explicit or implicit. Explicit feedback specifies an explicit rate, while implicit feedback indicates that congestion is either present or not present. The source might receive explicit and implicit feedback in the same RM cell. ABR connections have a minimum guaranteed rate that cannot be reduced by either explicit or implicit feedback. The means by which the feedback is used to help optimize bandwidth use is flow control.

Flow Control

In networks using LAN switches, the only flow control available is at the link level and is proprietary. It usually emulates flow control by buffering stop and go, which adversely affects performance. However, because flow control in an ATM network is a characteristic of a logical station in a virtual channel built end-to-end, it provides superior control at the virtual circuit level.

We would like to thank Kishore Jotwani, Gary Byrd, Andy Rindos, and Jill Kaufman for their guidance and support.

We would also like to thank Jim Glekas, Henri Sourbes, Haissam Alaiwan, and Joan Cavin for previous material they have published or prepared that was used in this document.

In addition, we would like to thank Tom Hadley, Bob Louden, Joe Robinson, Greg Pappas, Marty Cummings of Chrysler, Phil Emer of NCSU, John Parker, Brian Dorling, Rick Lacks, Bob Peckham, John Magri, Ed Wagstaff, Cordell Anderson, John Hart, Jim Kunkel, Dave Yeary, and Hugh Fish who gave generously of their time and considerable talents.

Also, we would like to thank Lori Phelps, who helped create and edit this print, as well as others some of whom are cited above for helping to proofread this document and provide valuable comments and corrections.

Special thanks to Maria Petrillo for her assistance in making my stay comfortable, productive, and enjoyable.

Finally, we would like to thank IBM's customers, especially those who are mentioned in this document, without whom we could not be successful.

© Copyright 1998, IBM Corporation