Welcome to the Power Users community on Codidact!
Power Users is a Q&A site for questions about the usage of computer software and hardware. We are still a small site and would like to grow, so please consider joining our community. We are looking forward to your questions and answers; they are the building blocks of a repository of knowledge we are building together.
Best practices for small, internet-critical company's ISP for maximal uptime
This is a small call center with maybe 20 employees. I do not know how many calls it is taking per day, how many field agents it is remotely dispatching, or how many customers it has, but you can estimate it is getting a five minute call once every ten minutes from 6am to 11:30pm at night (this is of course an average). These calls are from multiple states in the US. The calls are inherently connected to internet connectivity to input and retrieve important data regarding customers and jobs and such.
The internet should be fast and up ideally at all times. Recently we had our internet either go completely down or be unpredictably on and off for about 48 hours. Comcast Business / Xfinity said that there was some "scheduled maintenance" explaining the internet outage. We cannot as a business accept 2 days of scheduled maintenance, unfortunately.
Is there a very standard way in which a small internet-focused company can ensures a very high degree of internet uptime?
I called Comcast and they said all of their plans have the same uptime. They said normally small businesses get a plan from a second internet provider, for example AT&T, to reduce the chances of the internet going down. I am wondering if that is actually the best solution, or if it is an inefficient use of money.
What is the standard setup in this scenario?
3 answers
Generally speaking the best thing is two different internet providers. Different means not just who you send the bill to but different physical networks. In most areas, your options are:
- One cable provider (Comcast/Xfinity, Cox, etc.)
- One telephone company provider (Verizon, AT&T)
That's pretty much it for hardwired connections. Most other companies are piggybacking on a cable or telephone system for at least the "last mile", so the redundancy between that "independent" provider and the cable or telephone company is minimal.
There is however a third alternative that is rapidly becoming available: satellite. But specifically Starlink. The first few rounds of satellite internet were high latency (e.g., based on geosynchronous satellites), high cost, low speed or all of the above. Starlink is still being developed, and it is not perfect, but it generally has low latency, relatively low cost and fairly high speed (not as fast as the best cable or fiber, but fast enough for most typical uses).
If I had a need for a backup internet connection (primary being cable or telco) I would definitely go with Starlink.
Beyond simply having the connection available, there are some issues with routers and other details depending on whether you need automatic switchover or not and other factors. But those are the little details.
0 comment threads
Your only real option here is redundancy.
Besides what manassehkatz explained already, you can also open a second office in a different location. It seems unlikely that internet would fail in two different cities at once, and this will work even when there is an ISP monopoly in each area.
0 comment threads
[Comcast] said normally small businesses get a plan from a second internet provider, for example AT&T, to reduce the chances of the internet going down. I am wondering if that is actually the best solution, or if it is an inefficient use of money.
The typical solution to a situation where you cannot afford downtime is some kind of redundancy. Depending on exactly what you're doing with it, that redundancy can either carry you through an outage or allow you to plan repairs to the primary system for when those repairs are less disruptive to typical operations.
Redundancy of any kind is always an inefficiency. However, you have to weigh the cost of the redundancy against the cost of a failure; and then consider the probability of a failure.
If you "cannot accept" downtime that presumably you even are notified of or can find out about in advance (as tends to be the case with scheduled maintenance), then the only two remaining options for you would seem to be:
- Negotiate an agreement with the provider where they agree to not have downtime that impacts you. This appears unlikely to succeed and won't help you while the provider still does have an outage.
- Have an alternative means of accomplishing the goal (for example, Internet connectivity for VoIP routing) which is unlikely to suffer from any common-mode failure, to provide you with redundant service.
Starlink has already been mentioned. It looks like their service in the US currently starts at $120/month, which isn't even pennies on the dollar compared to the cost of having 20 employees (as a crude example, 20 employees working 22 8-hour days per month at only $10/hour is $35K/month). It seems likely that you're using VoIP, maybe even some kind of locally-terminated VoIP trunking, in which case the routing of the IP traffic shouldn't matter much, although you would need to check with your VoIP telephony provider to be certain.
Terrestrial mobile broadband can also be an option, but seems quite likely to be relatively saturated in a situation where there is a large outage for any reason. I also strongly suspect that by the time you have a large enough data allotment to carry you through a significant outage, the cost for that data plan will rival satellite service with relatively little to speak for it in practice.
Once you have a redundant means of connecting to the Internet, the issue becomes failure switch-over (can be automated, but is likely to disconnect any calls currently in progress; still less disruptive than being knocked offline completely) and how to keep that hardware from becoming a single point of failure (which is entirely doable at a cost in money and/or complexity). It is entirely possible to set things up such that a single-uplink disruption is effectively invisible to network users, but to fully benefit from that takes significant effort well past simply signing up for Internet service over a second medium from a second company.
As an aside, you really owe it to everyone involved to draft some kind of business continuity plan. One part of such a plan is to identify and ideally rectify any single points of failure that can negatively impact the ability of the company to conduct normal operations. If typical business operation is completely dependent on Internet connectivity, then anything which might disrupt Internet connectivity belongs there: everything from an accidentally-cut fiber bundle during public services repair work to your Internet provider messing up their core network routing configuration.
Incidentally, again, most of which are typically solved using, you guessed it, redundancy and minimizing common-mode failures.
0 comment threads
1 comment thread