User`s guide

Introduction
1-11
preserves network bandwidth for important user data and frees up
valuable end station processing. By defining virtual workgroups,
broadcasts will only be seen by other end stations within the same
virtual workgroup. With the functionality to define workgroups by
port grouping, IP network address and/or IPX network number, a
station can be part of multiple workgroups based on their location
and protocol.
Each workgroup can be defined by port, IP network address
and/or IPX network number. A total of 100 virtual workgroups
can be defined on each ATX LAN Switch. The ATX LAN Switch
can route between IP workgroups but all other workgroups will
need an external router (See Workgroup to Workgroup
Communication). For additional information, see section 3.8
Virtual Workgroup LCM Commands.
1.6.6 ATX Packet Processing Engine
The ATX architecture, diagrammed in Figure 1-2, is based on dual
29030 RISC processors on the Packet Processing Engine version
3(PPE-3). In addition, it includes the following:
• At least one RISC processor per i/o module
• Backplane providing 1.6 Gbps capacity, with a load balancing
architecture for maximum accessibility for I/O modules
• A 2mb shared RAM architecture, which is optimized using
adaptive buffer allocation. Adaptive Buffer Allocation (ABA) is
an algorithm providing a sophisticated distribution of packet
buffering to meet varied utilization demands per port.