T 2.45 Conceptual deficiencies of a network

Planning the design and expansion of a network is a critical success factor for network operations. Especially given the shorter and shorter innovation cycles in the field of IT, networks that cannot be adapted to new requirements due to their design may quickly become bottlenecks:

• Depending on the requirements determination of network subscribers (e.g. work groups) for the confidentiality of data and the integrity of the network, the network must be designed accordingly. Otherwise, confidential data of a work group may be read by other, unauthorised network subscribers. Seen from this angle, the confidentiality may also be lost by relocating participants of work groups or the entire work group if it is not possible to configure and/or change new confidential areas in the network. This threat affects the integrity of the network and/or the integrity of network segments in a similar way.
  
  Example: A separate sub-network isolated with the help of a router was configured for a work group with special requirements for the confidentiality and integrity of its data. This segment is restricted to one building due to the way the cables are routed. After several employees of this work group were relocated to another building, these employees have to communicate using the normal productive network. The confidentiality and the integrity of the data can no longer be guaranteed.
• If new applications are operated on the network which have higher bandwidth requirements than those taken into consideration in the planning phase, this may quickly result in a loss of availability of the entire network if it is no longer possible to sufficiently scale the network infrastructure as a consequence of conceptual deficiencies (loss of availability due to overload). Depending on the selected segmentation of the network, the loss of availability may also only affect individual segments of the network.
  
  Example: Networks which have grown according to demand, which can still be found frequently today, often contain backbone segments with a low maximum bandwidth for historical reasons, e.g. token ring or Ethernet segments. This restriction of the speed in the backbone segment affects the availability of the entire network in the event of a high additional load.
• Networks only suitable for connecting proprietary systems may also cause a loss of availability if systems not suitable for the network are connected to the network (loss of availability due to non-interoperable network components).
  
  Example: Networks that are not system-neutral can predominantly be found in mainframe environments for networking the mainframes with the related terminals. Frequently, these are networks installed for terminal or printer operation and not suitable for operating other architectures (e.g. Ethernet). This applies both to the cabling used and the active network components. If this is nevertheless attempted, the proprietary network will no longer be available in general. An option for integrating two architectures may include connection via a gateway.
• When using active network components not designed for the use of certain protocols, it may not be possible to use any additionally required services or protocols.
  
  Example: In a network exclusively consisting of active network components which only support IP routing or IP switching, no Novell Netware network operating system can be operated based on SPX/IPX.
• When using passive network components which entail a restriction to the network access protocols to be operated on them, it may no longer be possible to scale the network in the future.
  
  Example: No ATM may be used in a network designed exclusively using 50 Ohm coaxial cables. No 100 Ohm Ethernet components can be operated on networks designed with 150 Ohm twisted-pair cables. The conceptual deficiencies, which sometimes have historical reasons, among others, result in cost-intensive changes to the network infrastructure.

Networks may be designed in an application-, system-, or service-neutral manner, but due to a very heterogeneous component landscape they may require support that cannot be provided by the operating personal any more. This may cause a loss of the availability of the network if malfunctions or failures of passive or active network components cannot be eliminated quickly enough due to a lack of personnel resources.