FTTH Terminal Box
Field Assembly connector
EPON system
Fiber Switch
SFP Transceiver
Media Converter
PDH Multiplexer
Optical Automatic Protect
PLC Splitter
Optical Circulator
Optical Isolator
Optical Switch
Patch Cord/Connector
Optical Adaptor
Optical Attenuator
Patch Panel
Fiber Optic Closure
In-door Terminal Box
Optical Distribution Frame
FTTH Drop Cable
Optical Cable

¡ôSwitching and Routing (2)
  Evolving this type of network to Gigabit Ethernet speeds requires no major architectural changes. By simply adding bandwidth, the ¡°any-to-any¡± intranet traffic patterns are easily supported. New LAN switch features, such as IP Multicase ¡°snooping¡± techniques can limit the effort of multimedia multicast and broadcast traffic flooding the entire network.

But this approach has limitations. Since there is no inherent broadcast containment, the widespread propagation of broadcasts can waste precious bandwidth in large networks. Even worse, desktop CUP cycles are potentially wasted. Even worse, desktop CUP cycles are potentially wasted as each and station processes each broadcast. Large flat networks can also be difficult to implement. Additionally, organizations may not have the address space to implement a completely flat network. As networks grow, some level of layer 3 control is usually required

 Switch Where You Can, Route Where You Must

This approach leverages the simplicity of a flat, switched network, but implements routing where necessary for security, broadcast containment or address management. The primary design goal is limit the amount of traffic that requires routing in order to increase performance and reduce routing administration costs while maintaining wire-speed switching between end systems on the same logical subnet,

If traffic flows are predictable enough that the network can be designed to keep 70 to 80 percent of the traffic within specific logical groups, virtual LANs (VLANs) can be used to create these logical groups. The router¡¯s role is then too move traffic between the VLANs. This allows intra-workgroup performance to scale with additional bandwidth, while trading off some performance between workgroups as traffic passes through the legacy router. Traffic through the router can be further reduced by configuring serves to be members of more than one VLAN, giving multiple workgroups access to server without traversing a router,

However, some organizations cannot always predict and limit inter-subnet or inter-VLAN traffic. In emerging intranet environments, where users access information from many different servers distributed throughout the enterprise ,traffic flows are often unpredictable and unstable ,In such instance the router may become the bottleneck to network performance ,

Route Once, Switch Afterwards

The ¡°route once, switch afterwards¡± option is new, incorporating a variety of proposed techniques like ¡°IP switching¡± and ¡°cut-through routing¡±. The consistent theme for these techniques is that communication between end systems on separate subnets initially flows through the routs, with subsequent traffic ¡°shortcut¡± to the faster switched path. These techniques vary in how the shortcut is initiated. Some leverage the end stations, some use intervening or special switches or routers, others do some combination.

 The benefits of ¡°route once, switch afterwards ¡± approaches are clear: faster, high-bandwidth paths are used for both intra-and inter-subnet traffic. The scalability, cost and migration is one which works over existing switches and routers.

bc-optics (hk) limited copyright © 2006. all right reserved    design by sendnet.inc