Tuesday, January 28, 2020

Would It Have Made a Difference If Telcos Had Stuck with ATM?

It is no longer a question, but there was a time not so long ago when global telcos argued for asynchronous transfer mode (broadband ISDN) as the next-generation protocol, rather than TCP/IP.  

A report issued by Ovum in 1999 argued that “telecommunications providers are expected to reinvent themselves as operators of packet-switched communications networks by 2005.”
“Growth in Internet Protocol (IP) services is expected to fuel the transition,” Ovum argued.

Of course, all that was before the internet basically recreated the business, making connectivity providers a part of the broader ecosystem of applications, services and devices requiring internet connectivity to function. 

In retrospect, it might seem obvious that the shift of all media types (voice, video, image, text) to digital formats would make TCP/IP a rational choice. But that was not the case in the telecommunications industry from the 1980s to the first decade of the 21st century. Telco engineers argued that ATM was the better choice to handle all media types. 

But the internet, cheap bandwidth and cheap computing all were key forces changing the economics and desirability of IP, compared to ATM. 


Once internet apps became mass market activities, network priorities shifted from “voice optimized” to “data optimized.”

Connection-oriented protocols historically were favored by wide area network managers, while connectionless protocols were favored by local area network managers. The relative cost of bandwidth drove much of the decision making.

WAN bandwidth was relatively expensive, LAN bandwidth was not. That meant the overhead associated with connectionless protocols such as TCP/IP did not matter. On WANs, packet overhead mattered more, so lower header overhead was an advantage. 

It is by no means clear that the choice of a connectionless transmission system, instead of the connection-oriented ATM, would have changed the strategic position of the connectivity provider part of the internet ecosystem. Indeed, one key argument for IP was simply cost: IP devices and network elements were much cheaper than ATM-capable devices. 

One might argue the global telecom industry simply had no choice but to go with IP, no matter what its historic preferences might have been.

No comments:

Will AI Actually Boost Productivity and Consumer Demand? Maybe Not

A recent report by PwC suggests artificial intelligence will generate $15.7 trillion in economic impact to 2030. Most of us, reading, seein...