SEACOM downtime after local backhaul problems

[InTheCube]

Just a thought: If there are 3 redundant paths, then the reserve capacity drops to 3:2 (1.5:1), which would be a lot more cost effective. In this respect, the more cables the better.

No.. you are wrong. 3 redundant paths will cost more then having 2 redundant paths. Yes with 3-way redundancy, the cost of each redundant route comes down individually. But then when you add the cost of the 2 extra routes together, you are back at square one, in fact it will cost even more.

It is cheaper to buy (hypothetically) a single 1GBPS link on one route, then to buy two 0.5GBPS links on 2 different routes. The 1GBPS link will be cheaper because of bulk discounts/savings.

 

[Gordon_R]

No.. you are wrong. 3 redundant paths will cost more then having 2 redundant paths. Yes with 3-way redundancy, the cost of each redundant route comes down individually. But then when you add the cost of the 2 extra routes together, you are back at square one, in fact it will cost even more.

It is cheaper to buy (hypothetically) a single 1GBPS link on one route, then to buy two 0.5GBPS links on 2 different routes. The 1GBPS link will be cheaper because of bulk discounts/savings.

I didn't mean 3 redundant paths to guarantee 100% uptime, but that if 1 out of the 3 paths goes down, then the other 2 can handle the remaining traffic at a slightly reduced capacity. At present a 2:1 ratio means severe congestion as soon as a link goes down (assuming they are equal capacity).

It is likely that as each new cable comes online, an ISP will spread their options, by buyng additional capcity on each redundant network. I am sure that growth in demand in the next year or two will make this a reality, while maintaining the economies of scale.

 

[AndrewAlston]

P.S. I did notice that TENET had to switch to its secondary path yesterday (I didn't record when), but the downtime was quite short.

Hi Gordon, just a note, switching to secondary path on our network is not an indication of downtime at all. The switch was a test of a new redundant circuit going from Mtunzini Landing station to our Durban point of presence. You will see more such flips between the two circuits on Saturday night as well, while we continue to test.

These type of active vs protection circuit switches you see on our graphs can be caused by numerous events, and because we run a policy that says not to automatically revert to one particular circuit, but stick with what is working, you can see switches where it flips path and stays there for a few days at a time. But yeah, like I say, a flip on the active vs protection path does not necessarily indicate downtime or a problem on the network.