Both the peer-to-peer network operating system model and client/server network operating system model have seen deployment implementations using a these topologies.

While looking at the Direct Link Topologies we learned how they rapidly become disproportionately expensive once more than three devices need to be connected. Officially, this is a type of Mesh Topology, as you will soon learn. So let us to it starting with the various Ring Topologies.

The Ring Topology Network

With a Ring Topology network, each node (Servers, Workstations, and Peripherals) connects directly to both of its immediately adjacent nodes. The cable circles around the nodes connecting each in turn before eventually connecting to the first node again.

Now we have all nodes connected in one complete self-contained loop and this is the reason that this type of enclosed loop physical connectivity structure is a Ring Topology.

In the previous article entitled “Network Operating System Star Topology”, I discussed the topic of physically star-wired logical ring topologies. I also introduced IBM's proprietary Token Ring Topology. I shall add a little more as this is the best way of explaining the issues concerned with Ring Topologies.

Token Ring

IBM developed the Token Ring protocol in the mid-1980s. As the name Token Ring (see Figure TR 1) suggests a closed loop (or circle), wiring schema connects all member machines of the Token Ring network segment. This “ring” must remain intact for communications to be possible. If the “ring” breaks then no all machine-to-machine communication is possible.

In the previous article “Network Operating System Star Topology”, we learned that a ring topology is very difficult to install and configure. It also poses the issue of multiple single-points-of-failures. In order to address these issues IBM created a special network device they called a Multi Station Access Unit (MSAU). Now each node connects directly to a centrally placed Multi Station Access Unit (MSAU).

Externally the MSAU looks just like any other hub or switch. Internally at the hardware level however, the Multi-Station Access Unit (MSAU) contains wiring which allows information to pass from one device to another in a logical circle or ring. This reduced the multiple single-points-of-failures to just one, the MSAU.

Technically speaking IBM's Token Ring protocol uses a star-wired ring topology with the Token Ring Multi-Station Access Unit (MSAU) is the heart of the Token Ring universe. The signal travels from node to MSAU, then to the next node, then back to the MSAU, then to the next node and then back to the MSAU and so and on it goes.

Token Ring Network Media Access

Probably the most defining feature of the Token Ring network is the manner in which machines gain access to the transmission medium. The access method involved in a Token Ring network uses token-passing access.

A special “electronic token” (special frame) is circulated around the network. Each node will take this “token” frame from the media and if not wanting to transmit the node will place the “token” frame back onto the transmission medium.

When an empty token reaches a computer that wishes to transmit, the computer attaches the data to the “token” and then places the entire package (“token” and data) back onto the transmission medium.

The “token” with the data then circulates until it eventually arrives at the intended destination node, which then captures the data. If it needs to reply and can do so immediately it does so.

If not the computer that just received the “token” and data package, places an empty “token” onto the transmission medium and so the whole process goes on.

One of the biggest advantages of this system was that it made collisions impossible. There would never be two machines attempting to communicate at the once since only one machine could hold the “token” at a time.

Token Ring Topology Summary

• Transmission Media - Star-wired ring using twisted pair or fiber optic cable
• Transmission speeds of 4 Mbps or 16 Mbps
• In Token Ring Networks, the signal travels around the network from one computer to another in a Logical Ring
• Due to the increasing popularity of Ethernet, the use of Token Ring has decreased

Fiber Distributed Data Interface (FDDI)

FDDI is another ring topology but unlike Token Ring, it uses dual counter-rotating rings.

Tree Topology

A Tree Topology is really a hybrid topology as it combines the characteristics of both the linear bus and star topologies. The Tree Topology consists of a number of groups of star-configured workstations connected to a linear bus backbone cable. Tree topologies allow for the expansion of an existing network.

Tree Topology Advantages

• Point-to-point wiring for individual segments
• Supported by several hardware and software venders

Tree Topology Disadvantages

• Overall length of each segment is limited by the type of cabling used
• If the backbone line breaks, the entire segment goes down
• More difficult to configure and wire than other topologies

Mesh Topology

The mesh topology is all about removing Single-Points-of-Failure for internetworking components and infrastructure alike.

Mesh topologies achieve this by interconnecting all network components with multiple cabling connections. In this way mesh topology implementations aim to provide multiple points-of-connection for every network device to every other network device thereby providing a direct dedicated link between all pairs of network devices.

Whenever a device or communications link fails, the alternate pathways become active until the troublesome device(s) returns to active duty.

Mesh Topology Advantages

• Redundant Network Designs - Multiple Alternative Pathways
• 24/7 Availability
• Decreasing the number of Single-Points-of-Failure

Mesh Topology Disadvantages

• Cost - The extra required cabling contributes considerably to the cost of implementing a mesh topology
• Since all other network components in a mesh topology also require multiple duplicate network infrastructure devices (2 or more routers, switches, etc.) for every primary network device (routers, switches, etc.).
• Redundant network infrastructure devices contribute considerably to the higher startup costs of a mesh network in comparison to other topologies
• A complex system/network is always harder to construct and deploy than smaller more easily managed sub-networks
• Every device will need to have multiple NIC - More Additional Expense

Network Operating System (NOS) Images

As mentioned last time, I am having trouble sending, the graphics with these articles. Therefore, I have decided to include said graphics for the entire series in a separate photo gallery entitled “Network Operating System (NOS) Images”. I will be posting it in the next day or so.

I will also keep on trying to send them embedded into each article. Once the upgrade to the lines in our area is complete, things should be better.

Well, this completes this short discussion of network operating system topologies. We will now have a quick look at the special class of machines that run the network operating and use it to deliver services. You guessed it I am talking about servers. This will be the topic for the next article titled “Network Operating System Servers”. Until then enjoy!

The Star Topology Network

With a Star Topology network, each node (Servers, Workstations, and Peripherals) connects directly to a centrally located Hub / Concentrator (seeFigure ST 1).

Data always travels from the transmitting node through the hub or concentrator before continuing to its destination.

Network Management

It is the job of the hub or concentrator to manage and control all functions of the network. The central hub/concentrator also behaves like a repeater for the data flow in that it reshapes the signal before passing it on.

Switches do a similar job as well. Having discussed hubs, concentrators, switches and repeaters in earlier articles, I will not dwell on them here.

Star Topology Supported Transmission Media

One of the greatest assets of the star topology is that it can use practically any type of available transmission media. For copper-based wired networks, some flavor of twisted pair cable (usually UTP) is the most common type of transmission media generally found in star topology networks.

Copper-Based Transmission Media

Unshielded Twisted Pair (UTP) is the most abundant variety of copper cable available today. Therefore, it should come as no great surprise to discover that Unshielded Twisted Pair (UTP) cabling is in deed the predominant type of transmission medium (cable) found in the star topology networks of today.

This holds true for most of the other topologies to boot. Volume production has meant that UTP cable is considerably cheaper than its competitors are. Advances in UTP cable capabilities over the years have seen the continual evolution of UTP cabling to the point where it is today.

Star topology networks built around Shielded Twisted Pair (STP) and coaxial cable are still in active service today. They are disappearing quite rapidly however as existing networks upgrade to higher performing infrastructures. These “new” network infrastructures tend to be a mix of wireless and wired technologies.

The Wired Component

For the wired component Unshielded Twisted Pair (UTP) CAT5/CAT6 or above dominate. The use of fiber-optic cable in star topology networks is generally limited to those situations demanding “extreme” performance, immunity from Electro-Magnetic Interference (EMI), long cable runs and other scenarios where UTP is unsuitable.

Fiscal Considerations

One should never lose sight of the fact that above all else the greatest advantage that copper-based transmission media (wires/cables) has over its rivals is cost. Copper cable is considerably cheaper than fiber-optic cable.

Star Topology Protocols

Protocols commonly associated with a star topology include Ethernet and LocalTalk (Apple's Proprietary Protocol). Both protocols, Ethernet and LocalTalk require a different NIC

Star Topology Ad Hoc Networks

The star topology also lends itself for use in ad hoc networking scenarios and hence is very popular with small business and home networks. In fact, this sector constitutes over 60% of all networks. The driving impetus here has been the desire to save money by sharing an Internet connection between numerous machines.

Star topology is very popular with peer-to-peer network scenarios. The ease with which nodes come and go is one of the biggest reasons that the star topology is so popular with peer-to-peer networks. LAN parties and gaming sessions are a breeze. For these types of ad hoc network use, the star topology model is most definitely the easiest to install and configure.

Plug "N" Play Connectivity

Most ADSL, broadband modems on the market today include an integrated 4-port, switch capable of transparent bridging. The average consumer and small business alike find themselves ideally placed to use these types of “plug "n" play” network and Internet connectivity devices.

Star-Wired Ring Topology

A star-wired ring topology such as that implemented in IBM's Token Ring networks appears very similar to the classical star topology externally (physically). Both have nodes that connect to a central hub/concentrator or switch. The nodes themselves only differ in terms of the Network Interface Card (NIC) used.

• In order to connect machines to an Ethernet network an Ethernet NIC is required
• On the other hand, to connect to a Token Ring network a Token Ring NIC is required

Multi-Station Access Unit (MSAU)

All member devices of a network that uses a star-wired ring topology must connect to a special type of device called a Multi-Station Access Unit (MSAU). Externally the MSAU appears to be very much like a normal hub, switch or LAN Switch.

Internally at the hardware level however, the Multi-Station Access Unit (MSAU) contains wiring which allows information to pass from one device to another in a logical circle or ring.

IBM's Token Ring protocol uses a star-wired ring topology. The Token Ring Multi-Station Access Unit (MSAU) is the heart of the Token Ring universe.

Star Topology Advantages

Cable Selection - The star topology permits the use of UTP, STP, coaxial and fiber-optic cabling. This allows for greater diversity and flexibility at setup time. The star topology is backwards compatible with the older cable types so long as the central hub/switch supports it.

Cable Installation - Cabling is straightforward and easy to install particularly when UTP CAT5 or above is used.

Connectivity - With devices, connecting directly to the central hub/switch greater freedom of configuration is possible. Moving a PC to somewhere else in the room or even to another star topology room is simple and speedy.

Reduced Outages - Network disruption and downtime is greatly reduced when connecting or removing devices. Generally, there is no need to power down any device other than the device slated for removal.

Peer-to-Peer - Simplified ad hoc Peer-to-Peer (P2P) networking schemata

Troubleshooting - Easier to use & more streamlined troubleshooting courtesy of the ability to isolate problem devices made possible by the star topology's physical connectivity simplicity

Upgrading - Component replacement, upgrading, the installation of service packs, hot fixes, patches and the addition of new features is a lot easier with a star topology than linear bus and physical ring topologies

LAN Switches - Implementation of a star topology in conjunction with LAN Switches, gives us the ability to fragment collision domains. This offers great improvements in network performance over bus and ring topologies

Routers and Virtual Local Area Networks (VLAN) - Adding routers and Virtual Local Area Networks (VLAN) will also improve network performance as well as providing for greater security control over resource allocation and access.

Star Topology Disadvantages

More Cable - More cable is required to implement a star topology than would be the case for a linear bus topology

Single Point of Failure - If the central hub or switch dies then the entire network will be unable to communicate with each other & externals networks

More Expensive - Hubs & switches are not free and additional cable adds to initial outlay costs to the implementation of the star topology in comparison to the linear bus topology

Network Operating System (NOS) Images

As I have been having trouble sending, the graphics with these articles I have decided to include said graphics for the entire series in a separate photo gallery entitled “Network Operating System (NOS) Images”.

I will also keep on trying to send them embedded into each article. Once the upgrade to the lines in our area is complete, things should be better.

Next time in “Network Operating System Ring and Tree Topologies”, we will look into the Ring and Tree Topologies including their implementation indicators. Until then enjoy!

Dual Ring Technology

Fiber Distributed Data Interface (FDDI) evolved from the IEEE 802.4 token bus timed token protocol. Fundamentally, FDDI is a ring network similar to IBM's Token Ring network but with a number of crucial differences.

The most noticeable difference between the two is that a Fiber Distributed Data Interface (FDDI) network is a dual-attached, counter-rotating token ring topology. One ring is the primary transmission ring delivering transmission speeds of up to 100 M-bit/sec.

The other ring (secondary) is for backup. This means that the secondary ring is inactive and remains so for as long as the primary ring is functional. In the event that the primary ring fails, the secondary ring becomes active. Now all traffic goes to the secondary ring for transmission. Because of this built-in redundancy, FDDI is a fault tolerant technology.

Extra Network Transmission Capacity

Network administrators do have the option of determining that if they do not need to have the secondary ring idle for the event of primary ring failure backup they can use it for data transport thereby doubling the network's capacity to 200 M-bit/sec.

Frame Size

FDDI has a larger maximum-frame size than standard 100 M-bit/sec Ethernet. This means much better throughput than standard 100 M-bit/sec Ethernet.

Immunity to Electro-Magnetic Interference (EMI) and Noise (Cross Talk)

Because FDDI uses fiber optic transmission media, it is highly resistant if not impervious to the effects of Electro-Magnetic Interference (EMI). This makes FDDI the ideal candidate of choice in those situations where the installation environments are prone to high exposure levels of EMI. Fiber optic cabling is also less susceptible to many environmental variables such as water damage.

Coverage and Scalability

Being able to cover large distances is most definitely a major plus for FDDI. Another is that FDDI can readily scale to support thousands of users.

Recent Developments

Fiber Distributed Data Interface II (FDDI-II) has added support for circuit-switched services thereby enabling FDDI to carry both voice and video signals as well.

Work connecting FDDI networks to developing Synchronous Optical Network SONET networks has just recently begun.

Modified FDDI Physical Connectivity Practices

In order to retain much of the resilience delivered by the original physical connectivity specifications of FDDI while greatly reducing the setup cost of a new or expanded installation we are seeing physical connectivity modifications being adopted.

The most popular seems to be the spreading practice of not attaching all workstations to both rings. Instead, network administrators and engineers are dual connecting only a small number of infrastructure devices such as routers and switches to both rings and the workstations to the infrastructure devices. The result is still a resilient network but at greatly reduced cost of implementation.

Security Considerations

Because fiber optic cable is virtually “tap-proof”, FDDI definitely beats the competition in the reduction of security risk/threat exposure stakes.

FDDI Standards

ANSI X3.166-1989, Physical Medium Dependent (PMD) -- also ISO 9314-3

ANSI X3.148-1988, Physical Layer Protocol (PHY) -- also ISO 9314-1

ANSI X3.139-1987, Media Access Control (MAC) -- also ISO 9314-2

ANSI X3.229-1994, Station Management (SMT) -- also ISO 9314-6

ANSI X3.184-1993, Single Mode Fiber Physical Medium Dependent (SMF-PMD) -- also ISO 9314-4

Conclusions

Today however we find that due to current speed (since 1994 we have had the option of Gigabit Ethernet), cost, ease of installation, manageability & flexibility, other media (copper) are more popular than fiber optic cable.

Even though we now have 10 Gigabit/s capable fiber optic systems, the biggest downside of cost has an overwhelming negative buying impact upon the majority of the public.

Considering all factors, there can be little doubt that FDDI will continue to be the technology of choice for a good many case specific applications (niche) such as ISP high capacity, high speed, high availability, and maximum-security network data backbones.