Following up on my previous post [1], let’s take a look at the pure 100G wavelength option. Right now they can do pure 40G waves and they think they can use what they learned getting to 40G to get to 100G on a single wavelength. From a network design point of view, this is quite desirable because everything new is inside the black box, all you have to do is buy the new black boxes. However, it has taken 40G some 8 years to become affordable relative to 10G, and even then just barely. If one is going to scale that solution by tens and hundreds without reducing prices at the same rate, it becomes unaffordable quickly.
Advantages:
- fiber-poor carriers can increase capacity forever
- the paradigm doesn’t change, network design looks the same as you scale
Disadvantages
- each jump requires a technological leap of faith
- just because you figure out *how* doesn’t mean you can do it affordably in the same time frame.
Implications
- There are no real losers because the basic infrastructure that everyone has will still be just fine.
- The winners are those with limited fiber, because that limit no longer matters – they can always keep up. Cogent, for instance, is a big fan of this solution. So are VZ and AT&T, who have mostly older fiber and would have to dig in order to install new stuff at some point.
Complications
- If it requires new breeds of optical fiber to follow this route, then the eventual winners are those with empty conduits – Level 3 and Qwest.
2 Comments To "The road to 100G and beyond – pure 100G wavelengths"
#1 Comment By Ralph Doncaster On June 25, 2008 @ 8:46 am
I’ll make a wild guess and say that it will be at least 10 years after we have a 100G standard before there is a standard for terabit links over fiber.
100G is more than the current transit capacity of most countries in the world.
[2]
100G is enough capacity for 5000 20mbps ATSC video signals, and video seems to be the main driver of internet traffic growth. In the coming years it will make sense to use 100G long-haul links into the 10 biggest US cities (1+M population), but what about the 250 cities with a 100K+ population? Until the cost of 100G approaches 2x the cost of 10G it just won’t make sense.
-Ralph
#2 Pingback By The road to 100G and beyond – bandwidth aggregation On June 25, 2008 @ 8:01 pm
[…] time we looked at pure 100G wavelengths, now let’s look at bandwidth aggregation. In order to reach 100G why not take ten 10G […]