Do you agree with the checklist and best practices mentioned below? Do you know any that should be added to these lists. Let me know and I’ll get them added if they are pertinent.
Network Health Checklist
1) Ethernet segments should not have a sustained utilization of 40% or higher.
2) All Ethernet segments should be switched. No shared segments(hub-based).
3) No WAN links should feature a sustained utilization of 70% or higher.
4) WAN response times should be generally less than 100ms.
5) LAN response times should be around 2ms.
6) No segments have more than one CRC error per MB of data.
7) Segments should be no more than 20% multicast/broadcast traffic.
8) Device CPU utilization should not exceed 75% over 5 minute intervals.
9) Output queue drops should not exceed 100 in an hour.
10) Input queue drops should not exceed 50 in an hour.
11) Buffer misses should not exceed 25 in an hour.
12) Ignored packets should not exceed 10 in an hour.
Access Layer Best Practices
1) QoS for performance.
2) Redundancy for availability.
3) Limit VLANs to a single closet.
4) RPVST+ instead of STP or PVST+.
5) DTP set to desirable/desirable. ** I do not agree with this one **
6) VTP transparent mode.
7) Disable trunk mode on access ports.
8) Routing in the access layer.
9) Portfast on edge ports.
Distribution Layer Best Practices
1) Aggregate bandwidth together into EtherChannels.
3) Security mechanisms.
4) FHRP protocols.
6) Address summarization.
7) Triangles and not squares for redundancy in and between layers.
Core Layer Best Practices
1) Redundant triangle designs between switches.
2) Fast switching at Layer 2. No routing if possible.
3) Multilayer switches.