Design The Network - Mitel Inter-Tel 5000 Installation And Maintenance Manual

Design the Network

For optimum network operation, follow these guidelines:
•
•
•
•
•
•
•
•
•
•
•
®
Inter-Tel 5000 Installation Manual – Issue 2.4, May 2008
Carefully manage the traffic on a central node: The central node is susceptible to
unintentional bottlenecks. Tips for preventing bottlenecks on the central node are:
o If the node contains an EM or Voice Processing Unit voice mail system with more than
16 ports and more than 128 Executive Display, Professional Display, or Model 8560
endpoints, redistribute the load by moving some of the Executive Display, Professional
Display, or Model 8560 endpoints to another node, creating a new node, if necessary.
o If a networked node with an EM or Voice Processing Unit with Automated Attendant
and/or Call Routing Applications is expected to handle more than 1000 calls in an
hour, use a separate EM or Voice Processing Unit to provide the announcements.
o If a node with a centralized voice mail supports more than four nodes or a total of 1500
off-node devices, redistribute the load by reducing the number of endpoints on the
central node to 200 (adding another node, if necessary).
o A node used as an outgoing call center should not have more than 100 agents.
Redistribute the load among several nodes or place the call center on its own node.
o A node should not have more than 3 ISDN Primary Rate trunks for private networking
or to the central office. Redistribute the load by moving ISDN trunks to another node,
adding a node, if necessary.
o For very busy systems or large networks, equip the satellite nodes with their own local
and long distance trunks where possible.
o If the central node is used for all long distance calls, make sure it is equipped with
enough trunks to handle the peak traffic.
Limit the maximum number of display endpoints to 206 per node: Redistribute the
load by adding another node if necessary.
Eliminate data errors whenever possible: Currently, the system can experience
slowdowns during error dumps. The most common preventable errors are:
o T1 errors
o Fiber errors
o Glare: Ensure the Inter-Tel system and central office are programmed to minimize
glare by initiating outgoing calls on the opposite end of a trunk group from the starting
point for incoming calls. Even when configuring in this manner, glare is possible when
the system gets very busy, but the likelihood of a glare condition is greatly reduced.
Turn off diagnostics except when working with Engineering or Technical Support to
troubleshoot a problem. Diagnostics require processor resources and they can adversely
affect the speed of heavily loaded nodes.
Ensure that enough T1 spans are installed between nodes to handle the traffic: In
most cases, a single T1 span between nodes will suffice. However, if increased call
capacity anticipated between two nodes, add another T1 span to handle the traffic.
Avoid doing DB Programming during peak business hours.
Avoid using System OAI on a heavily loaded node: Redistribute the load by moving
traffic not associated with the System OAI application to another node (adding a node, if
necessary).
When using Paging: Avoid using a lighted DSS/BLF Button for the main page zone—for
example, 9600 or 9601.
Limit page zones to 60 or fewer endpoints for customers who need to make more than 30
pages per hour. Use external paging whenever possible.
Avoid overloading all-ring hunt groups: Avoid all-ring hunt groups with 10 or more
endpoints that are expected to handle more than 500 calls in an hour. Use linear or
distributed ringing instead of all-ring.
Avoid large hunt groups: Avoid creating hunt groups containing more than 100
members. Distribute the hunt groups over several nodes, if possible.
Appendix A: Inter-Tel Private Networking
Design the Network
Page A-5