Monday, November 29, 2021

Redundant Internet & UPS home worker equipment.

Hi all,

I am looking for hardware / ecosystem recommendations for the following:

Scenario

About 10 home workers in different countries, not all have reliable power or internet connections.

Device(s) Requirements:

  • Primary Internet Connection (Ethernet/ADSL2+) with 4G Backup, auto-failover
    • Depending on the connection we may need just a Ethernet WAN, or a modem as well. Mostly likely the WAN would plug into existing customers modem, but looking at options.
    • Same with the 4G, This may come in the form of Integrated 4G, WAN2 or USB 4G modem.
  • Cloud Management (Need to be able to configure / fix this remotely, Plug and Play to get online from end-user perspective)
  • Metrics (Dynamic DNS and SNMP absolute worst case, graphite export or something similar would be better).
  • USB for UPS monitoring
  • Wifi and LAN Port Out.

Metrics & Alerting

I am looking to centralize the alerts, monitoring, etc ideally into something like Grafana, or if the ecosystem can do this for us, that is also fine.

  • Monitor the status of a wired & wireless backup internet connection (I need to see if the backup goes down before there is a primary failure).
  • Monitor the status of a UPS system. I want to see when there is power and when not how much backup time is available.
  • Both Real-time alerting is required as well as viewing historic data to notice any trends (e.g How long do power outages last on average, do we need a bigger UPS, how reliable are the ISP's, latency monitoring would also be useful)

I have played around with pfsence and tunnels, with snmp, etc. I can make it work, however there is too much technical overhead required for management, lack of centralized management cloud and complexity for end user setup.

Meraki and Ubquity seem like ok options, but I really don't know enough about what better solutions may be available, or within these specific brands what combinations would work best. I am also not looking at spending thousands per deploy.

Thanks in advance.



No comments:

Post a Comment