1-4
8237 Hub Installation and Planning Guide
Stacks
A
stack
is formed when up to ten 8237s are connected together. To create a stack,
two or more 8237s must be interconnected using the Stack Link cable. For more
information on interconnecting 8237s, see "Interconnecting 8237s" on page 1-5 and
"Connecting Hubs with Stack Link Cables" on page 2-5.
A
managed
stack is made up of any combination of Model 001s, Model 002s, and
Model 003s, as long as at least one Model 002 or Model 003 is included in the stack.
In stacks with only one Model 002 or Model 003, that unit will automatically be
designated the primary management agent, regardless of its physical position in the
stack. For stacks having more than one Model 002 or Model 003, the following rules
are observed:
•
In a stack with a Model 002 and Model 003 together, the Model 003 will
automatically be designated the primary management agent.
•
Among the same models in a stack, the unit that is physically nearest the top of
the stack will automatically be designated the primary management agent and
the remainder will be designated backup management agents in order of their
physical position in the stack (when connected as described in “Connecting
Hubs with Stack Link Cables” on page 2-5).
•
In a stack with more than one management unit, in case of the failure of the
primary management agent, the management agent that is physically next
lowest in position in the stack will automatically assume primary management
agent responsibility.
For more information on managed stacks, see "Understanding Managed Stacks” on
page 1-6.
An
unmanaged
stack is made up exclusively of Model 001s.
There are four ways to manage the 8237:
Using a local VT100 terminal emulator out-of-band
Using SNMP over SLIP out-of-band
Using Telnet over an IP network connection in-band
Using SNMP over an IP or IPX network connection in-band
Interconnecting Typical Ethernet Hubs
The best way to appreciate the benefit of stacking hubs is to examine the method of
interconnection used by simpler Ethernet workgroup hubs, which interconnect by
cascading
. Cascaded hubs are in the same collision domain. Figure 1.2 on page
1-5 shows three Ethernet workgroup hubs labled A, B, and C. Hubs B and C are
cascaded from Hub A.
Because data sent from the workstation to the server must pass through three hubs
along the way, the server is said to be three
repeater hops
from the workstation. A
repeater hop is counted whenever an Ethernet frame passes through a repeater.
The IEEE 802.3 standard specifies that a frame sent from one workstation to
another should not pass through more than four repeaters on the way to its
destination.
If there are six ports on Hub A, you can cascade up to four more hubs from Hub A
and still have a maximum of three repeater hops between any two workstations. A
limitation of cascading, however, is that ports that could be used to attach
workstations are being used to interconnect hubs.
Summary of Contents for 8237
Page 1: ......
Page 2: ......
Page 7: ...vi 8237 Hub Installation and Planning Guide...
Page 15: ...xiv 8237 Hub Installation and Planning Guide...
Page 41: ...2 8 8237 Hub Installation and Planning Guide...
Page 91: ...3 50 8237 Hub Installation and Planning Guide...
Page 109: ...4 18 8237 Hub Installation and Planning Guide Figure 4 20 8237 Field Replaceable Parts...
Page 113: ...4 22 8237 Hub Installation and Planning Guide...
Page 117: ...A 4 8237 Hub Installation and Planning Guide...
Page 147: ...B 30 8237 Hub Installation and Planning Guide...
Page 166: ...Part Number 85H8847 Printed in Taiwan 85H8847 GA27 4186 00...