Lan Topologies Essay, Research Paper
When designing the layout of a local area network, an organization can choose from a variety of
different technologies on which to base the backbone of their LAN. The technology choices that the
organization adopts will play a critical role in the performance of its LAN. Two important LAN
backbone issues, which I will introduce, include:
? A network based on Ethernet or Token Ring technology
? Factors to consider when choosing a Network Operating System
LANs are also known as subnets, which are single networks. Overall, LANs use a hybrid
TCP/IP-IEEE (Institute of Electrical and Electronic Engineers) framework consisting of 6 layers:
application, transport, internet, logical link control layer, media access control layer, and physical
layer. The IEEE has many standards committees to develop standards for LANs. The most
widely known is the 802.3 Working Group, whose Ethernet standards dominate LAN technology.
Another is the 802.5 Working Group, which created the competing Token-Ring network
standards used in many organizations. At the media access control layer (MAC), CSMA/CD
(Carrier sense multiple access/collision detection) governs communication between stations.
Ethernet uses a bus transmission, which is characterized by broadcasting. One station transmits,
and this transmission is broadcasted to all other stations. In Ethernet 10Base-T and 100Base-TX,
the hub does the actual broadcasting. In contrast, Token Ring network stations connect to access
units, which are connected in a ring. A special frame called a token signals around the ring and
stations which grab hold of the token may transmit. Whereas Ethernet uses CSMA/CD for the
MAC layer, Token Ring networks use token passing to determine when each station may transmit.
The growth of today’s LAN traffic is pushing network administrators to look too higher-speed
network technologies to solve the bandwidth problem. Highly reliable networks are critical to the success
of the business organization, so ease of installation and support are primary considerations in the choice
of network technology. Network administrators today have several alternatives to choose from. Although
each network faces different issues, Gigabit Ethernet meets several key criteria for choosing a high-speed
? Easy, straightforward migration to higher performance levels without disruption
? Low cost of ownership—including both purchase cost and support cost
? Capability to support new applications and data types
One of the most important questions network administrators face is how to get higher bandwidth
without disrupting the existing network. Gigabit Ethernet follows the same form and function as its 10
Mbps and 100 Mbps Ethernet precursors, allowing a straightforward migration to higher-speed
networking. All three Ethernet speeds use the same IEEE 802.3 frame format, full-duplex operation and
flow control methods. In half-duplex mode, Gigabit Ethernet employs the same fundamental CSMA/CD
access method to resolve contention for the shared media. Simply stated, Gigabit Ethernet is Ethernet,
only faster. Because of the high scalability of Ethernet technology, it is simple to connect existing lower-
speed Ethernet devices to Gigabit Ethernet devices using LAN switches or routers to adapt one physical
line speed to the other. This evolutionary upgrade path allows Gigabit Ethernet to be seamlessly
integrated into existing Ethernet and Fast Ethernet networks.
Network administrators today face a myriad of internetworking choices and network design options.
They are combining routed and switched networks, and building intranets of increasing scale. Ethernet
networks are shared (using repeaters) and switched based on bandwidth and cost requirements. The
choice of a high-speed network, however, should not restrict the choice of internetworking or network
topology. Gigabit Ethernet can be switched, routed and shared. All of today’s internetworking
technologies, as well as such technologies such as IP-specific switching and Layer 3 switching, are fully
compatible with Gigabit Ethernet, just as they are with Ethernet and Fast Ethernet. Gigabit Ethernet is
available in a full duplex repeater (with the accompanying low cost per port) as well as on LAN switches
Today, Ethernet networks are rapidly approaching the reliability level associated with their telephone
ancestors, and are relatively simple to understand and administer. The development of Gigabit Ethernet
extends the scalability of Ethernet even further. Now Ethernet scales from 10 to 100 to 1000 Mbps. There
has been a rapid overall decrease in the per-port price for Fast Ethernet and Ethernet products, and for
shared hub ports, the difference between their costs is narrowing. Over time, Gigabit
Ethernet per port costs can be expected to experience similar price decreases.
First introduced in 1985, the Token Ting evolution got underway in 1989 when the first
16mbps Token Ting adapter arrived. Token Ring switching later arrived in 1995 providing high
speed backbone connectivity between rings and enabling servers to benefit from dedicated full
duplex 16 Mbps connections. In his 1998 research paper “High Speed Token Ring”, Martin
Taylor explains that Token Ting switches provide high speed backbone connectivity between
multiple 4 or 16 Mbps rings. Servers can be connected directly to the switch, benefiting from a
dedicated 16 Mbps link, which eliminates the need for shared bandwidth. These dedicated links
can run in full duplex mode, providing a clear transmission channel in both directions at the same
time without noise.
However, since the maximum speed currently supported on Token Ring is 16 Mbps,
orgainzations are forced to look to other technologies if they want a solution for “inter-switch”
links and for high speed server connections. Higher speed technologies that are available today to
provide a solution for inter-switch links and server connections in Token Ting environments
include Fast Ethernet and ATM. Another alternative, as suggested by Taylor in “High Speed
Token Ring”, is High Speed Token Ring (HSTR) itself.
HSTR is needed primarily to address the problems of inter-switch links and fast server
connections. Since Token Ring standards already provide a model for full duplex dedicated
Token passing with a maximum speed of 16 Mbps, Taylor argues “all we need to do is to adapt
this to run at a higher speed.” To keep costs low (such as the competing Ethernet technology),
HSTR will be based on the new solution of an existing and widely available high speed
transmission technology, such as 100 Mbps Fast Ethernet or 155 Mbps ATM. The HSTR
solution is based on the IEEE 802.5r standard for Dedicated Token Ring, adapted to run over the
100 mbps physical transmission scheme, which is used by, dedicated Fast Ethernet. There will also
be a Gigabit Token Ring solution, based on IEEE 802.5r adapted to run over the Gigabit Ethernet
Today, Network Operating Systems (NOS’s) have gone far beyond their roots of file and print
services. Other functions, such as communications, database, application, and management services, have
become equally important in business environments. Companies using NOS’s have found supporting
large numbers of single-function servers to be costly and complex. These companies have started
consolidating to a smaller number of larger, more powerful, multifunction servers. This trend makes it
imperative for the NOS to have the flexibility to support such configurations. NOS applications vary a
great deal in their use of system features and functions, NOS functions are implemented two ways:
1. As a standalone operating environment that may or may not allow the support of additional services, such as database or electronic messaging.
2. As an additional service layered upon a general-purpose operating system, such as Unix or OpenVMS.
In a International Data Corporation (IDC) white Paper, by Dan Kusnetzky, Kusnetzky argues
that stand alone NOS environments provide the highest performance when the only requirements are file
and print services. They often are less stable and robust when supporting other functions. When
multifunction servers are required, layering NOS functions on a general-purpose operating system
becomes the best choice. Kusnetzky addresses the following issues concerning Network Operating
? Architecture — Companies need to be concerned with the architecture of a NOS. Does the NOS support multiprocessor systems? Are these systems supported asymmetric or symmetric multiprocessing systems? Can NOS functions be partitioned to run on more than one processor simultaneously? Does the operating system support multiple microprocessor architectures?
? Scalability — Companies expecting a NOS to support small, medium-sized, and large environments using the same NOS need to consider the following questions. What is the minimum and maximum memory, disk cache, and disk supported by each NOS? What is the maximum number of file locks, open files, concurrent clients, servers in any domain, and domains supported by each NOS.
? Availability and reliability features — File locking, and various types of client failover when a server fails are important to companies that rely on the network as a critical part of their day-to-day business operations.
? Clients supported — Most companies have a broad array of desktop devices. A NOS must support all of the desktop devices being used in a company if it is to be a viable candidate for adoption. Typical devicesthat must be supported are systems running DOS, DOS/Windows, Windows for Workgroups, Windows 95, Windows NT Workstation, Macintosh, OS/2, and Unix.
? Network printing — Printing is one of the primary functions of a NOS. When selecting a NOS, companies must address the following questions ahead of time:
How many printers are supported per server?
Can multiple printers be driven by one print queue?
Can multiple print queues drive one printer?
Will the NOS deliver an alarm message to an operator if a printing problem arises?
Will the NOS inform the user when a print job is completed?
Can the print function be managed remotely?
? Network media — Many different network media types are in use in companies today. To be successful, a NOS must support a broad collection of network media, including all ethernet media, all token-ring media, asynchronous and synchronous telephone lines, packet-switched data networks, fiber-optic, and Integrated Services Digital Network (ISDN)..
? Network protocols — Companies selecting a NOS must be certain that the protocols on the corporate network are compatible with the NOS. These protocols may include AppleTalk filing protocol (AFP), TCP/IP, Telnet, Simple Mail Transfer Protocol (SMTP), Simple Network Management Protocol (SNMP), SNA, File Transfer Protocol (FTP), Internetwork Packet Exchange/Sequenced Packet Exchange (IPX/SPX), NetWare file services, NetWare print services, NetBIOS, and NetBEUI.
? Network services — NOS platforms should provide the functions needed to support large business environments, including support for directory services that allow users to access network services without having to know the network address, and security services that control access to directory management functions.
? Security — Companies need to feel comfortable that corporate data assets are secure. This category
evaluates a NOS platforms’ standards compliance (B2 and C2 security ratings), support of access
control lists, disk quotas, automatic discovery of and management of intruders, support of callback
modems, support of security management systems, and whether or not encryption services are
? Application development tools — NOS platforms are increasingly expected to support application
services as well as the traditional file and print services. However, a NOS should support reviews
third-generation languages (3GLs), fourth-generation languages (4GLs), and also object-oriented
development tools for applications developments.
? Data access — As NOS platforms are expected to support more complex applications, they are expected to support file access methods and indexed file access methods. This category evaluates NOS platform support of these file services.
? Database support — Database software is often part of the infrastructure of a distributed application.
The NOS platform should support various database software supplied by different vendors.
? Applications — This category reviews which application services are available on each NOS platform. Application services include mail client, mail server, word processing client, spreadsheet client, integrated office suite, or Lotus Notes. Companies selecting a NOS platform for both file/print services and application services should consider NOS platforms with a strong ranking in this
Thus, two different types of technology (Token Ring vs. Ethernet) and the functional requirements
Demanded from an efficient Network Operating System have been discussed. Cost and scalability seem
to be the two main concerns that organizations have when looking into these technologies. In writing my
executive summary, I interviewed Monesh Vargheese, a network engineer at the UIC Medical Center to
evaluate these technologies. Vargheese explained the UIC Medical Center was recently converted from
Token Ring to Ethernet. “With the emergence of Fast Ethernet (Full duplex, auto sensing 10/100 Mbps
NICs) technology, the old Token Ring technology no longer added any value to the network”, according
to Vargheese. According to Vargheese, a reason for the switch to Ethernet was speed. With Token Ring,
the transmission speed of 16 Mbps for any particular workstation was shared among all other
workstations within that specific ring. With the new Ethernet technology, workstations can now transmit
at 100 Mbps and these NICs costed as much as the 16 Mbps Token Ring NICs. Cost and scalability were
the primary reasons behind the move to full Ethernet. Scalability is an important factor because the
network engineers plan on upgrading to Gigabit Ethernet in the near future and Fast Ethernet allows this
without any major changes to the network. However, Vargheese did tell me that there was less
maintenance with Token Ring and fewer collisions when the network was Token Ring. Vargheese also
explained to me that the Network Operating System they chose to implement was Windows NT Server
4.0. This was because most of the applications used are made to run on NT, according to Vargheese.
From the research and the interviews that I have conducted, it seems that Ethernet is the better choice
over Token Ring because of its lower cost, higher scalability, and speed. As for Network Operating
Systems, a NOS that addresses issues such as: Architecture, Scalability, Number of Clients supported,
Network printing Network protocols, Network services, Security, and Database support needs to be taken