My Internet Service went down yesterday and I had to revert to my backup provider.
Network Outages due to upstream provider failure are endless…
And even if your provider is not to blame, there are endless hackers out there instigating DDoS attacks , some with an ax to grind others just for random entertainment.
Although this sampling of news stories is not very scientific, I could literally spend a month clipping these articles. There are new ones every day, and that is just the major ones that get reported. If I informally poll our customers, almost every single one of them has seen a DDoS attack of some kind in the past year, and all have had some sort of upstream Internet outages within the last couple of years.
Now if I ask how many have had critical Network Equipment go down due to hardware failure, that list shrinks to maybe 1 or 2 percent of our customers. Basically, what this tells me is you have a 100 percent chance of a Network outage for some period of time every year due to a problem upstream with your provider. You have a 2 percent chance due to a hardware failure with your local core Router/Firewall/Bandwidth/Switches.
To put that another way, for every 50 outages caused by external events at your provider beyond your control, you have 1 event due to internal hardware failure.
The solution is to have multiple distinct Internet Providers on hand at all times, so if one goes down you can switch over to the other. As I said there is nothing wrong with the idea of sourcing redundant local equipment, but statistically it is much more important to get a second Internet provider sourced before investing in redundant equipment.
Here is another article highlighting the prevalence network outages.
Note: Although DDoS attacks are provider Independent, your chances of stopping or mitigating the attack are enhanced by having multiple providers.
Other causes of failures:
Yes, wireless topologies are notoriously unstable, and so are applications running on Web Servers, both of which can cause service outages to local users. These types outages are usually not on the same order as catastrophic hardware failure problems or upstream failures. Outages with wireless equipment and service are usually related to these products getting into a bad state, and are not associated with a complete loss of communication to the outside world. You’ll still need to re-boot these systems to get them back into a good state.