MEC (Mobile Edge Computing)

Telcos Need App Partners to Make Edge Profitable

Mobile edge computing will fail to generate profits for telcos unless they can forge closer ties with application developers, according to a leading professor of computer science at the prestigious Carnegie Mellon University.

Mahadev Satyanarayanan, who has previously worked with technology players including Intel Corp. (Nasdaq: INTC) and IBM Corp. (NYSE: IBM), said telcos risk losing another battle to IT giants if they cannot make edge computing more "friendly" to app developers.

"You have to play an active role in the creation of new apps to make the infrastructure worthwhile," said Satyanarayanan during the Mobile Edge Computing Congress held earlier this month in Munich. "If you don't, you won't make the profits -- the giants of the IT industry will."

The comments come as telcos weigh up the pros and cons of deploying IT resources in access networks, where they are much closer to end users, instead of relying on more centralized data centers. (See Will ETSI Lose Its Edge as Fog Rolls In? and ETSI Gets Edgy About Mobile.)

Such "edge computing," as it has become known, promises operational cost savings and performance improvements for telcos and could ultimately give them a competitive advantage over cloud providers with more centralized facilities.

But doubts persist over the business case for edge computing. For one thing, the costs of rolling out a nationwide edge-computing network are likely to be astronomical. Instead, operators are initially looking for specific scenarios in which the economics would stack up.

Persuading developers to come up with new and compelling applications for edge computing could obviously provide a much-needed boost for the emerging technology. Yet telcos have previously struggled to build effective links with the software and developer communities.

According to Satyanarayanan, they should start by coming up with technology and financial incentives to convince developers that edge computing is a viable alternative to the cloud.

"The preferred approach will be to run it in the cloud because that is an easier path for developers," he says. "You have to make the edge friendly to app developers and this requires something new from telcos."

Want to know more about cloud services? Check out our dedicated cloud services content channel here on Light Reading.

According to Satyanarayanan, one option is to pump investments into technologies such as machine learning, computer vision and the Internet of Things that would make the edge appear more "valuable" from a developer perspective.

As far as he is concerned, edge computing could hold various attractions over the cloud if developed in the right way, including much lower latency and the ability to support real-time analytics. The edge could also be developed as an alternative to the cloud in the event of an outage. "If connectivity to the cloud is broken, the ability to have a fallback service at the edge will be very important," says Satyanarayanan.

Carnegie's remarks are especially interesting given ongoing efforts by European Telecommunications Standards Institute (ETSI) , a major specifications group, to develop a range of open APIs [application programming interfaces] that would attract the interest of developers.

"To enable new markets we need developers to come up with exciting things people want to use," says Adrian Neal, the vice president of ETSI's industry specifications group for mobile edge computing. "If APIs are designed with industry best practice in mind, developers will find it easier to create code that can be ported on to MEC [mobile edge computing] servers."

Perhaps the biggest risk for telcos is that web-scale giants like Google (Nasdaq: GOOG) and Facebook -- which obviously maintain links with developers of all kinds -- seize the edge computing initiative. (See Telcos Give Vent to Edge-Computing Anxiety.)

Although neither currently maintains much edge infrastructure, both are becoming increasingly active in the access networks business and have the financial clout to support edge-computing investments.

— Iain Morris, Circle me on Google+ Follow me on TwitterVisit my LinkedIn profile, News Editor, Light Reading

MordyK 9/28/2016 | 7:54:38 PM
Re: Fragmentation issues Carol,

the approach of a carrier acting as a gateway interface for multiple carriers over a common API structure - akin to wholesale M2M roaming accounts - could work in theory, although I just can't see it actually happening.

So far when it comes to these sorts of initiatives, we see alot of talk at events about collaboration, But when it comes down to actually making these things happen, it has a way of disappearing from the radar until the next big telco shift conversation.
cnwedit 9/28/2016 | 7:15:00 PM
Re: Fragmentation issues It seems like telecom operators could do a better job of at least defining some common interfaces. That might be a critical first step to being easier companies with which to work. 

Because otherwise, you are right, few if any telecom operators will be a substantial enough market to attract developers in droves. 
MordyK 9/28/2016 | 4:03:11 PM
Fragmentation issues Iain,

The need for telco's to get their ducks in a row to make them attractive to developers is on the money, but the olution is not as simple as he makes it sound.

Webscale companies benefit from a "write-once, run everywhere" benefit that carriers simply do not have as they are regional and/or play in specific limited markets. If telco's wanto play they need to create a cenntral clearinghouse structured organization, where they can compete with a "write-once, run everywhere" environment and provide the added benefits - like edge capabilities - that only telco's can provide.

The banks recognised this in the 60's and created unified transaction networks like Visa and MasterCard, which kicktarted the electronic transaction market. So if telco's wanto play in the developer space they need a cetral developer interface of some sort.

To date, the last successful cooperation with telco's bringing out a mass market product is the venerable SMS/MMS.

Cooperation on NFV & SDN along with the open source work in this space, is a rare and very recent bright spot.
Sign In