Analytics/Big Data

Level 3 Does Big Data Differently

Most telecom activity in the data center space is associated with companies delivering cloud services -- such as Verizon, CenturyLink, AT&T and Windstream. Level 3 Communications, which opened its newest North American data center this week in Herndon, Va., also has a major data center footprint -- with 350 facilities globally -- as part of its own cloud play of a different kind. (See Unknown Document 710421.)

In addition to offering a video cloud service via its content delivery network, Level 3 Communications Inc. (NYSE: LVLT) is delivering infrastructure-as-a-service to help enterprises collect and analyze big data, as well as provide the backbone connectivity for many of the cloud connections.

In a conversation with Light Reading, Paul Savill, SVP of product management at Level 3, pointed to Level 3's ability to help businesses collect and aggregate the large and small streams of information that make up big data, while also connecting them to cloud computing resources such Amazon Web Services Inc. , for the computational power to do the analysis. His company is seeing significant growth in the enterprise sector because of the growth of big data.

Starbucks, a Level 3 customer, is a good example because it is creating the ability to report a plethora of detailed data about individual stores' activity, down to the temperature of the coffee in their pots, all of which can be stored and analyzed.

Our NFV & the Data Center event digs deeper into how telecom service providers are evolving their data center strategies. You can check it out here on Light Reading.

Level 3's networks help businesses collect all that data from individual locations in a cost-effective way and, through the Cloud Connect service, gain access to the computational power needed to analyze the data. (See Level 3 Offers Private Links to Cloud and Level 3 Cloud Connects Digital Realty.)

"The basis of big data is creating these reams of information that has great potential locked into it if you can put the right analysis against it and understand the correlations of what it is telling you," Savill says. "We don't have a big data service and we're not selling consulting services around big data or analytic processes. But we can play in that whole equation."

Both the network and data centers factor into Level 3's solution. For Starbucks, Level 3 had to "get creative" to produce a cost-effective network option for every store, and for a lot of small streams of data that have to be collected, Savill says. Only in aggregate does all of that become "big data" and have analytical value.

Level 3's ability to deliver cost-effective connections and aggregation of all these smaller streams of data will be greatly enhanced when the tw telecom deal closes, if it does so as expected, because tw telecom has a much bigger metro network footprint and three times as many office buildings on-net, he notes. (See Level 3 CTO Jack Waters: Network Integration Guru.)

But Level 3 is also leveraging its global reach, and here, Savill cites the example of a movie studio producing an animated movie through the collaboration of many different work groups, in different locations, who each work in their own time zone and return the work-in-progress to the cloud when they are done.

"We have customers in the media and movie industry who may be working on a project for a year -- an animated movie -- and have various locations where the work is done," he says. "Modern animation requires a massive amount of computational power to render a one-minute scene, and it's much more cost-effective to do this work on a cloud computing platform where they pay for that by the hour."

By owning its own network of data centers, Level 3 can provide places to store and aggregate data, and then the cloud connections to enable the analysis, through its Cloud Connect offer, which reaches AWS and Microsoft, and will be connecting to other large cloud operators, Savill says. The combination of ubiquitous networks and data collection sites is what the carrier is selling.

"As companies are moving their IT operations to the cloud, big data is one of the use cases that is pretty popular," Savill says. "The computational power that is needed to perform big data analysis can be very large and very sporadic."

Rather than invest in the cost of owning and managing servers that aren't needed for regular use, enterprises look to the cloud for on-demand compute power on a pay-what-you-use basis, he says.

"We are seeing companies take terabytes of data and push it to an off-site platform, perform the data analysis and pull the data back and shut down the cloud service," Savill says. CloudConnect lets them do that and only pay for the network services they use.

The Level 3 exec doesn’t commit his company to never offering a cloud computing service -- that may be in the cards at some point, he says -- but for now, providing the data center capacity and the network connections is a significant part of the carrier's focus on growing its enterprise services.

— Carol Wilson, Editor-at-Large, Light Reading

Page 1 / 2   >   >>
thebulk 8/21/2014 | 8:54:52 PM
Re: Good fit @Ariella, that's a great concept and I can see the benefit, but I'm sure it freaked a lot of customers out.
DHagar 8/21/2014 | 8:18:43 PM
Re: Good Fit @Carol Wilson, great blog.  I like your identification of the "sweet spot", being ubiquitous networks with data collection; I think that is what is attracting the multiple players.

I agree with you and Daniel that Level3 could decide to move into that space, but need to develop the proper foundation if they choose to do so.
cnwedit 8/21/2014 | 6:20:24 PM
Re: Good fit Daniel,

They may reach that conclusion themselves -- certainly they have the infrastructure for such a service. But it also requires management and software - at least that's what CenturyLink says in talking about their new private cloud offer.

Level 3 isn't saying they won't offer cloud services, that's just not the approach they are currently taking.
cnwedit 8/21/2014 | 6:20:22 PM
Re: Good fit Daniel,

They may reach that conclusion themselves -- certainly they have the infrastructure for such a service. But it also requires management and software - at least that's what CenturyLink says in talking about their new private cloud offer.

Level 3 isn't saying they won't offer cloud services, that's just not the approach they are currently taking.
Ariella 8/21/2014 | 4:41:24 PM
Re: Good fit @DanielC you don't care for SaaS? 
danielcawrey 8/21/2014 | 4:36:09 PM
Re: Good fit I would think that Level 3 would be incredibly competent in offering private cloud services to its customers.

That's really in my opinion the next generation of cloud technology. Too many companies today are relying on SaaS products/services that are not as secure as they really should. 
Ariella 8/21/2014 | 2:56:42 PM
Re: Good fit @thebulk it's possible some would be, but some do go even further, matching loyalty cards with faces and online offers. I wrote about that last year  here.
thebulk 8/21/2014 | 2:50:53 PM
Re: Good fit I am 100% with you on that, I just think there would be some minor opposition to such a program and it would likely be vocal enough to scare off many retailers.
Ariella 8/21/2014 | 2:46:31 PM
Re: Good fit @thebulk but they still know what you buy if you use the loyalty cards. I know of people who say they refuse to use them for that reason or who try to game the system by switching with friends. To tailor offers -- say coupons for diapers to those who buy baby food -- they use that data already. But they can also use the data in a general way to know what people tend to buy together to, say, plan a better layout for the store. 
thebulk 8/21/2014 | 2:38:09 PM
Re: Good fit @Ariella, normally I would say yes, because I love efficient systems. However; we have to take I to account the privacy concerns if the customers. I still remember how upset some were in the US when some stores first rolled out a loyalty card program. For that reason and that reason alone I think it's best to keep two seperat systems in place.
Page 1 / 2   >   >>
Sign In