Monday, April 8, 2019

Can 5G Providers Sell QoS?

Can 5G service providers charge a premium for low-latency performance guarantees, when the stated latency parameters--best effort--are already so low? That is a question that also might be asked in other ways.

Will 5G best effort service be good enough, latency-wise and bandwidth, to obviate the need for any additional quality of service features to preserve low latency and bandwidth?

Is there a market for quality of service when delivered bandwidth rates are so high, and latency performance so much better than 4G? In a broader sense, as network performance keeps getting better on both latency and bandwidth dimensions, can connectivity providers actually sell customers on QoS-assured services?

Also, some would argue, it becomes problematic to try and maintain QoS packets are encrypted at the edge. A service provider cannot prioritize what it cannot see. And that is the growing trend as most traffic gets encrypted.  

By about 2020, estimates Openwave Mobility, fully 80 percent of all internet traffic will be encrypted. In 2017, more than 70 percent of all traffic is encrypted.

The other change is the emergence of edge computing for latency-sensitive applications. We can assume that the whole point of edge computing is to provide a level of quality assurance that cannot otherwise be obtained.

As content delivery networks provide such assurances to enterprises and content suppliers for a fee, so it is likely that edge computing networks or other networks relying on network slicing to maintain low-latency performance will be sold as a service to enterprises who want that latency protection.

Such deals do not violate network neutrality rules, which do not apply to business services such as content delivery networks. So, ultimately, between encryption, network slicing, edge computing and CDNS, there might actually not be much of a market for consumer services featuring QoS.

Best-effort-only never has been part of the vision for next-generation networks, whatever might have been proposed for the public internet. According to the International Telecommunications Union, “a Next Generation Network (NGN) is a packet-based network able to provide services including Telecommunication Services and able to make use of multiple broadband, QoS-enabled transport technologies and in which service-related functions are independent from underlying transport-related technologies.”

No comments:

Costs of Creating Machine Learning Models is Up Sharply

With the caveat that we must be careful about making linear extrapolations into the future, training costs of state-of-the-art AI models hav...