The main goal of the web presence is to teach concepts and help accelerate the implementation of cost effective, reliable IT. However, I have carried many hats in my life and seen many needs.
The Need
As a Network and Systems Engineer I know how complex systems can get and how many moving parts there are to make everything simply function and how adding redundancy triples that complexity instead of doubling it.
As a Network and Systems Administrator I know how hard it can be to track down an issue. Is it a VLAN, cable cut, the ISP is running slow, the solar flares simply causing wifi issues? There's a bucket of issues.
As a Software Developer I know that I can do something about it.
Initial Discovery
I've worked with as many as 8 different manufacturers of managed switches in the same network. It gets frustrating, particularly if you are chasing down a VLAN or STP issue. And, if you're trying to have redundancy, you will have STP issues. I also worked at a position in a company that had a decent amount of money to spend.
I'm working on a solution to use SNMP (version 1, 2 or the more secure 3) to view and manage the complete network. This will allow you to see all of the paths in use, monitor the amount of data flowing, block/unblock ports, name them, change VLANs, etc.
Because SNMP is a standard that not all switches follow correctly, the automation will vary from environment to environment but here's a summary of the discovery features:
- You define at least the initial network or discover with local IP/subnet information.
- Discover SNMP and WMI devices in the environment via ICMP.
- Pull Routing/ARP information to see devices that haven't yet been discovered and networks that might need discovery.
- Pull Interface data via SNMP or WMI, including performance data. This should include STP, PoE, VLAN, and basic interface data.
- Use STP data to determine the head of the network and where redundant links are. If you aren't configuring STP this will undoubtedly be the wrong switch.
- Use LLDP, CDP, and Switch MAC address tables to determine how everything connects.
- Use the connecting information to determine if an unmanaged or an undiscovered switch exists.
- Provide the user the ability to manually link or override linkages. In some cases, merging "switch" data.
Ongoing Monitoring
There are a few ways that ongoing monitoring can occur out of the box.
- SNMP Traps can be used to be notified of switch changes.
- Syslogs can be ingested to keep aware of what is occurring on the switches. This is the least standardized but most effective method. Only switches of the types I have access to, or am provided data from, will be processed this way.
- SNMP Discovery scans can be run at regular intervals to see if anything has changed.