serves them right
They've got beaches, sunshine, string bikinis, carnival and football. Pah! We've got pubs and rain and...cheaper internet. No contest.
IP transit prices are still declining globally, but in bad news for 'net users of all kinds, the declines are slowing. In what could be an indicator of continuing heavy demand for transit services, research company Telegeography says what was once a rapid fall in transit prices has eased considerably. The company says the 50 …
So, for just over $1 per Mbps, you can get a port in London which covers the globe, probably with a service level guarantee about packet loss and latency. Or, for £50 per Mbps, your ISP can get connected to BT's exchanges, with no such guarantees.
I wonder if these prices levelling out might signify growing demand and investment coming back? There was a huge overbuild early in the .com bubble years, with everyone rushing to lay fibre - then the bubble burst and the backbone providers could buy up lots of fibre cheaply, pushing prices down for years afterwards. (You weren't paying to lay new fibre any more, just paying to light up a spare bit of the existing stuff.) I wonder if Telegeography or similar track that aspect: are the point to point links that make up transit networks getting tighter too?
I think yes and no. While prices and margins are low, investment is harder. But demand is still growing, primarily due to video. There was a lot of competition to build to common, core locations like the classic FLAPS (Frankfurt, London, Amsterdam, Paris, Stockholm) to pick up Euro routes. So bankruptcies and consolidation lead to a glut of cheap capacity on those routes. Technology moved on, so we've gone from 'expensive' SDH links to commodity Ethernet. Aggregation costs dropped in line with using cheap 1/10G, 10/10G, 10/40G or 10/100G switches with backbone links moving from 1>10>100G although a lot are still running at Nx10G.
Challenge for ISPs is funding those upgrade steps. When 10G first launched, router and ports were expensive. Costs fell pretty rapidly as volumes ramped up in the ISP sector, but more importantly in the enterprise and datacentre space. I don't think it'll happen so fast with 100G as the demand is much lower, ie far fewer enterprise customers really need 100G, and 100G costs are still high. Especially if you need to forklift upgrades to existing switches & routers, and your racks are full. Or the new kit needs more power.
Then on the fibre side, continued cost reductions assume fibre is available, but some providers who've been selling cheap dark fibre have sold out. There will also be issues with the age and type of fibre laid during the .com boom given fibre aging, and some fibre types aren't suitable for running the latest DWDM systems to drive Nx100G services.
But some providers are looking at pretty substantial investments to keep up with demand where market expectation seems to be continuous price reductions.. So for those providers, interesting times are ahead.
Isn't this department of the bleeding obvious time again*?
Low price = low room for movement
High price = lots more room for movement.
i.e. prices of real things tend to get less elastic as they approach zero. Especially if there are some underlying base or floor costs.
*Disclaimer - Im not even sure what a Transit price is.
Whilst Gordon 10 is correct, 'Low price = low room for movement', I think this is more to do with lack of capital investment and the barriers to entry being high.
If you are an existing player there is no incentive to put more capacity into the network as this will increase supply faster than demand resulting in lower prices and smaller ROI.
Unless there is a new disruptive technologie that reduces costs/barriers or the current capacity is used up then I think we will see monopoly/cartelisation and resulting price increases/service stagnation.
Prices on the expensive routes will fall as the players concentrate on the high margin routes, prices on the cheaper routes will stagnate until the infrastructure falls over and the Gov has to step in.
Biting the hand that feeds IT © 1998–2021