Pros/Cons Dual Best Effort ISP vs Fiber/MPLS.
-
@Dashrender said:
I'm just not sure I can make that jump with them... again, soft science, so it's pretty hard to really know.
Soft science is one thing. Completely disbelieving math is quite another.
-
@Dashrender said:
I know that on average we have 3 outages a year at each remote location, so assuming that to be the case, I hope and anticipate that to be what happens in the future. So now I consider options to mitigate those 3 failures, which is the purchase of a second line.
In this quote you list three outages per year as an average. But you did not list the length of those outages. Fifteen minutes or a full day are wildly different.
And remember that two overlapping services should reduce the outages by something like 10,000:1.
-
I'm not saying that the expensive fiber plus cable option is the wrong decision, I'm simply saying that the numbers provided don't add up to that. If $1K/hr of downtime is the real cost, that does not make the fiber line a good idea. The fiber has to save you from nearly one hour every month compared to the two commodity lines to pay for itself. And that's assuming that the slower speeds won't incur you any lost value (maybe they won't, so I can't add in that number.)
But given the factors that you have presented here, the fiber line is clearly ruled out as part of the equation from a business / technology / finance perspective. Only an emotion or political reason would justify it.
-
Dual lines are a bit like RAID 1. If you think about a single super high reliability SAS drive (you know, the $1,000 a drive kind) and you ran with no RAID, you would expect data loss about once every six to eight years. Just how things are. You might get lucky and make it fifteen years, you might be unlucky and lose data in a week. But you would be very normal to get six to eight years.
But if you went to RAID 1 and moved to commodity drives (WD Red or even Green) for $250 total, you would expect to get more than 160,000 years before experiencing data loss.
Because when you are dealing with redundancy of this type, where the frequency of failure is incredibly low and you have unconnected, overlapping coverage, the chances of both things failing at the same time before the first one that failed is repaired, the risks drop by a staggering degree.
And unlike the RAID scenario where any overlap equals full failure, the WAN scenario allows for soft mitigation where even in the unlikely event of an overlap, the redundancy would reduce the length of the outage by some unknown amount.
-
@coliver said:
@Dashrender said:
@scottalanmiller said:
So honestly, I think he's being reckless and illogical - just throwing away money because he's using emotion, not logic, to deal with the ambiguity of not having the details necessary for you to do your decision making.
Details only for the sake of conversation were as follows,
Company looses $1000/hr of downtime (all other loses are actually deferred income, because they would be rescheduled).
Company grosses 12 million a year, employees 80 people. The save of $8600/year moving to pure commodity is but a blip on that radar, and 'to them' isn't worth the potential risk by moving to a pure commodity setup.Wait... so $12,000 lost on the "commodity" connection, which is really only 12 hours over 7 years or less than two hours a year. You are looking at ~$2,000 of downtime a year vs the additional ~$8,000 of the cost of the reliable line? Just trying to make sure I understand.
Edit: Maths
nah, 3 outages per year. so $12,000 per year loss
-
So each outage is averaging four hours for $4K per outage?
-
@scottalanmiller said:
So each outage is averaging four hours for $4K per outage?
Yes, and even that is on the long side.
-
Great - Speaking to you guys and to a few others I know who work in IT management agree that there is no real advantage to keeping the fiber.
My night is going to go so much better now!
They thing they kept throwing in my face is that $8600/yr compared to the 12 Mill in gross profits is such a small blip that we would be foolish to not keep what we know is working.
We can tolerate downtime, we're not one of those crazy places that 'thinks' we can't tolerate downtime. Sure the BOD will be grumbling during one, but as you mentioned Scott, with the dual ISP setup the risk is equal to (but really probably better) than the fiber we have today.
-
@Dashrender said:
They thing they kept throwing in my face is that $8600/yr compared to the 12 Mill in gross profits is such a small blip that we would be foolish to not keep what we know is working.
What if your net profits are only $8,000 though? The difference could be the difference between profits and loss. That you have $12m gross profits means nothing at all. It's your net profits that would tell you if the number is significant.
-
Even small, rural farms (as opposed to urban farms) gross in the millions of dollars. But profits are generally tiny.
-
@scottalanmiller said:
@Dashrender said:
They thing they kept throwing in my face is that $8600/yr compared to the 12 Mill in gross profits is such a small blip that we would be foolish to not keep what we know is working.
What if your net profits are only $8,000 though? The difference could be the difference between profits and loss. That you have $12m gross profits means nothing at all. It's your net profits that would tell you if the number is significant.
I know I tried to point that out to them... they weren't hearing me.. I'm over it now.
-
@Dashrender said:
@scottalanmiller said:
@Dashrender said:
They thing they kept throwing in my face is that $8600/yr compared to the 12 Mill in gross profits is such a small blip that we would be foolish to not keep what we know is working.
What if your net profits are only $8,000 though? The difference could be the difference between profits and loss. That you have $12m gross profits means nothing at all. It's your net profits that would tell you if the number is significant.
I know I tried to point that out to them... they weren't hearing me.. I'm over it now.
Easy ways to spot people making emotional arguments and not actually looking out for the business. They might think that they are, but really only finance and IT are really supposed to be trained to think about the factors at play.
-
It's important to note that their point about the "known" reliability is valid. But they extrapolated from that one datum to an illogical conclusion based on it.
-
@scottalanmiller said:
It's important to note that their point about the "known" reliability is valid. But they extrapolated from that one datum to an illogical conclusion based on it.
Yeah that was the only part of their discussion that I did agree with - but I couldn't make the leap that it was worth staying there just for that, unless I expect the commodity connections to be so horrible as to be considered a risk not worth taking - which of course it currently is not.
-
I sent a request to Sales at Peplink today.. 8 hours later and still no reply ... I guess I'll have to call in the morning.
