TekSavvy → Re: Dear Marc...

Nitzguy, I'm not sure where you got that impression from but that's just not true. Aggregated or not, it's the exact same setup.

I was basing it on the CRTC tariffs published for connection to each POI....guestimating I guess because I don't know all the variables, but if you had to connect to each POI individually there would be an inherent cost to that (inital setup cost + ongoing monthly cost to maintain each link) would there not?

....Then if its not true, why drag heels and not just move to aggregated POI so you can serve the whole foot print?

If it is the same setup, you'd think that would have been done a long time ago, no?

I was guessing it was a cost issue since it would make no sense to remain dis-aggregated since you have to move to an aggregated POI eventually anyways.....

Sorry if I was completly wrong on this one...just figured it was a $$$ issue.

It's a dollars issue.
Any ISP on a multi-POI strategy has sunk bundles of money into providing backhaul - money which Rogers will NOT credit toward moving into aggregated POI. So the indie opts to try to return as much as possible on their sunk backhaul costs before moving to aggregated POI. In some (perhaps many) instances, the indie may not breakeven on the individual backhaul circuit costs prior to the CRTC mandated cutover to aggregated POI.

The difference between non-aggregated and aggregated POI is the router at which the TPIA traffic is routed off Rogers network ... Aggregated eliminates a few routers at the head end for TPIA traffic ... but it moves them to the aggregated POI ... and makes them huge routers instead!

*IF* Rogers had opted to taken the route of 'Unnaming' hiving off the customer and network IPs onto a "holding" company so that way users wouldn't know from the IP that a customer was Rogers or a TPIA customer, from the IP, and administered all the IPs that way, only identifying them internally, then the problems faced in node splits at the head ends would not have happened.

What you're talking about is dynamic routing. That's the only difference.

they have dynamic routing for themselves... and static for anybody who's not them. Ip's are assigned via DHCP in any case.. but assigning one of ours relies on a static mapping to exist to know how to get to our routers...

on their end, if they pull a card and remap a node to another part of the network, the dynamic routing just picks up on that and knows how to get out.. for us there's an extra step of making sure the statics are all correct.. removing them from where they were.. and adding them where they now are after a network change.

So will switching to aggregated stop congestion during peak hours?

It very well could depending on the type of congestion as the the aggregated links are purchased 10gbps at a time I believe. This creates a far larger pool of potential connectivity for the traffic to go through.

If you are encountering congestion at the local node level that is dependent on Rogers upgrading those areas, doing a node split, or you using a D3 modem.

You have summarized the tech point well! But from my understanding there is also a huge difference in billing model between aggregated and non-aggregated POI.

My understanding is that Teksavvy is paying a flat monthly fee by fiber in the non-aggregated model, while the fee is by 100 mbit/s increment in capacity in the aggregated model. From the many discussions I have seen on the forum, the non-aggregated model is much more affordable for TSI with the current pricing model, so making the switch is not strictly a tech problem. Also there might be long contracts on those non-aggregated POI fibers, this also drives costs of switching up.

The short answer is a definite *no*.

Congestion can happen at many levels.. to the degree that we have enough capacity at the POIs and at our data-center.. any congestion you might have would be at a lower level.

If we were aggregated, the same would still be true except that the capacity at the POIs would not be managed by us anymore...

The only benefit I can see is that by not having to manage the capacity at the POIs, we no longer have to worry about single gig links.. and only the big 10gig links at our data-center. That reduces the amount of work we need to do.. all we would need to do is to say we want to use 5gig out of the 10gig link and they would bump up the ceiling.. and that would be that. internally they would worry about the rest...

the flip side to that though is that we would then be measured and billed at their capacity based billing rate approved by the CRTC late last year... today, to flip over (assuming we didn't have all these contracts for the gig links at the POIs that each have multi year commitments to them that we would have to pay to get out of) would be more expensive on day one. In a year from now, it would be a lot more expensive.

Think about that. we're paying full retail rate now for those links. to convert, we would pay more. that means we'd be paying more than full retail rate... and other incumbents are way higher then them still... just goes to show how obscene those rates are.

...so, no. all aggregated is really doing is taking away responsibility from us and putting it on the incumbent to manage that part of the network. oddly enough, the CRTC wants us to take more responsibility.. but that's another story too.

does that answer your question?