How great are the advantages of infrastructure edge computing that is widely distributed within a single metro area? Not so clear, says Raul Martynek, DataBank CEO.
To be sure, 5G dramatically reduces the latency of the air interface, dropping round trip times from an average of 50 milliseconds to 60 ms (or greater) to sub-10 ms, Martynek says.
“With 5G, most cloud data centers will be 25 ms to 50 ms away from end-users, a significant improvement over 4G,” he argues. Deploying five smaller sites within a single metro area might improve round-trip latency by one to two milliseconds, he says.
“Once an application is deployed in a single location in a given market, the actual latency to reach any eyeball in that market is dramatically reduced compared to the common East/Central/West configuration of web scale data centers and the incremental benefits of micro data centers evaporate,” when 5G access is deployed, he argues.
So for many applications, requiring round-trip latency of three ms to five ms, for example, a single metro data center might work.
“It seems logical to us that before the large cloud and content players deploy at 10,000 cell tower locations, they will first deploy a single cluster in a traditional data center in the top metro markets that they are looking to service and be able to reach any eyeball in those geographies with very low latency,” Martynek says.
Such improvements obviously are relevant for any discussion of whether network slicing, for example, might allow creation of quality-assured customized networks with guaranteed levels of service (latency, packet loss, bandwidth, for example).
Many believe a new opportunity to supply quality-of-service guarantees on virtualized, 5G and other networks will exist. The counter argument is that routine levels of service will be so much better that it will be hard to convince most customers to pay for the higher-cost QoS-assured tiers of service.
That might especially be true for consumer customers.
Some might argue that 5G and better networks, growing competition for internet access services, plus content encryption, have killed the means and the demand for quality-assured consumer broadband services.
Even if ISPs wanted to sell services that prioritize quality, the technical ability to do so, and the consumer demand, are not present.
Ignore for a moment the politics of network neutrality. It can be argued that internet service providers rapidly are losing the technological ability to “degrade” or “slow down,” much less identify, packets they deliver. And without packet visibility, it is impossible to apply the quality of service mechanisms net neutrality proponents fear.
Keep in mind that any attempt to categorize and apply service level features to internet content becomes impossible when the data is encrypted. QoS packets are encrypted at the edge, by the app providers themselves.
When that happens, service providers cannot prioritize, because they have no idea what actual class or category of content is delivered. In other words, they cannot “tamper with what they 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.
There are other reasons packet discrimination is not possible, for technical and business reasons.
Can 5G service providers charge a premium for low-latency performance guarantees, when the stated latency parameters--best effort--are already so low? Could they charge a higher fee for faster speeds, when faster speeds are the norm?
That, essentially, is among the implications of fast 5G networks and metro-level computing.
No comments:
Post a Comment