x
Optical/IP

Smart Is in for 2012

1:40 PM -- This is the time of year when everyone starts compiling lists of what's been hot for 2011 or what's expected to sizzle in 2012. Verizon Communications Inc. (NYSE: VZ) annually gets a jump on this process, releasing what it thinks will be hot for business and for consumers. (See Verizon Forecasts Telecom Trends and Verizon Projects 2012 Consumer Trends.)

A lot of the trends are exactly what you'd expect -- more cloud, M2M, video and personalization on the business side; more social networking, video and user-generated content on the consumer side.

But there are two trends that are strong on both lists that I think warrant notice for very different reasons.

Both the consumer and business lists note the line is blurring between what is work and what isn't, as more consumers use their smartphones for both work and personal tasks.

It's the trend that led Verizon, among others, to devise ways to let enterprises piggyback their apps onto their employee-owned devices while protecting the business data should the employee leave or lose their phone. (See Verizon to Launch Enterprise App Stores and AT&T App Enables Work/Play Divide.)

This trend has been happening for some time, but smartphones are exacerbating the situation. That doesn't mean we have to like it, and I'm not sure I do. The fact that we can work everywhere at any time doesn't mean we should. But it's hard to blame service providers for making it easier, especially if they can sell it to enterprises as a feature.

More positively, the other trend making both lists is smarter networks, and this is where service providers should be concentrating their efforts. Every other network trend -- personalization, M2M, more video, that home/work combo -- depends on making networks smarter.

Network intelligence will also enable service providers to monetize broadband wireless more effectively by creating service tiers and packages that are more user-friendly and less onerous. Defining that network intelligence is not an easy task because it doesn't involve just one thing -- it's not just policy management and DPI, for instance.

I asked Roberta Mackintosh, Verizon's director of product marketing for solutions and professional services, how Verizon is defining "High IQ Networks" and her answer boiled down to tighter connections between networks and smart devices, network services and the cloud (i.e. data centers and data); and better signaling within the network itself, as well as smarter billing.

Basically that means flooding the network with intelligence but tying together the smart bits and pieces into a manageable piece. Your perspective on control of that piece will likely depend on where you are within the network infrastructure.

That's why two separate research projects can produce similar ideas about how policy is going to be deployed going forward, but reach very different conclusions. (See Pure-Play Policy's Future Is Fading and Policy Is Still Strategic, But Changing.)

And it's why other recent surveys tie smart services to higher profits. (See Smarter Services Pack Revenue Punch.)

So I'm boiling down my list of what's hot -- and good for you -- in 2012 to that one thing, a most uncontroversial choice: Networks and services have to get smarter.

— Carol Wilson, Chief Editor, Events, Light Reading

cnwedit 12/5/2012 | 4:48:50 PM
re: Smart Is in for 2012

I get your point. Unfortunately, research these days tends to be funded and that funding tends to come from vendor sponsors.


But I still think most of the research passes the "smell" test. Using network smarts to create services that are more tailored to customers so that they pay more for them.

paolo.franzoi 12/5/2012 | 4:48:50 PM
re: Smart Is in for 2012

 


I would like to see a study from somebody who is neutral about the smart thing leading to higher profits.  Having a vendor who is heck bent on doing that commissioning a study to show that seems rather self-serving in the result.


seven


 

cnwedit 12/5/2012 | 4:48:49 PM
re: Smart Is in for 2012

Again, you make a good point. But they have to add some intelligence to do tiered service and metering, as everyone has moved away from the all-you-can-eat model.


So the question isn't, will they add intelligence to the network, but how will they add it and how will they use it, beyond basic tiers?

paolo.franzoi 12/5/2012 | 4:48:49 PM
re: Smart Is in for 2012

Carol,


Where I think this may go wrong is on the cost side.


Imagine this scenario (and I recognize it is a hypotheical).  To get that 10% bump in revenue, AT&T has to replace every router in their network.  Good Deal or Bad Deal?  By the way, they have to sole source their entire routing network since none of these devices work in a common way.


seven


 

paolo.franzoi 12/5/2012 | 4:48:48 PM
re: Smart Is in for 2012

Well, see the thing is Carol...actually they don't need to add intelligence to the network for Metering or Tiered Service.


The counters are there already for any reasoned setup.  The question is are they going to do DPI and apply dynamic policy and pricing per protocol.  On top of that, how will they defeat anonymizers and other tricks to do this and make a lot more money. 


The problem with the Intelligent Network is that the people designing those things are not generally assuming a mix of equipment between vendors and the intelligence being an increment over existing equipment.


Let's use Tellabs for an example (who coined that study).  Will their product work in an environment where the customer has already deployed an entire packet core and aggregation network based on other devices?  If not, then how much cost is there to replace that entire network before $1 of new revenue can be added.


seven


 

paolo.franzoi 12/5/2012 | 4:48:43 PM
re: Smart Is in for 2012

chris,



How does this lower capital costs since basically in your description every piece of equipment in the network needs to be replaced?  Maybe I misunderstood your statements.


seven


 

chris-o-five 12/5/2012 | 4:48:43 PM
re: Smart Is in for 2012

Carriers are interested in Network Intelligence for two basic reasons, lower captial & operating costs and monetization of the network.  


The road to network intelligence has forks in it, Network vendors tell carriers that they need to have a homogenious set of gear across their network in order to best exploit the opportunities.  I on the otherhand, beleive it's all about the data (I would, I work for a Data Warehouse company).  


Let me explain, if the carriers objective is to provide the best "experience" on the network for the least cost, then network intelligence has to be about understanding how a customer's experience is impacted by the network.  To to this I, perscribe to the notion that the data has to be collected from the network, including devices, and brought together in such a way that it looks homogeneous and merged with information that the carrier knows about its customers.  Sounds like magic, it is sort of; it's the magic on data integration and I see it happen every day.  


From this point various users within the carrier's organization can exploit it to better manage the experience which in turn squeezes out costs and provides for those opportuniites to sell more services and retain customers.

paolo.franzoi 12/5/2012 | 4:48:42 PM
re: Smart Is in for 2012

 


Okay, well then what you are suggesting exists today and is done today.  That is capacity planning.  How much bandwidth do I need on average based on traffic loads.  That is what 15 minute bucket statistics are about.  In that mode, it does not matter what devices are talking or what they are talking to.  Only that the overall traffic has some set of characteristics.


When the carriers talk about smarts what they are talking about is per device usage and service usage.  So that they can bill people more money.  Something like:  For 1 Gbyte of web surfing you pay $1.  For 1 Gbyte of Youtube videos you pay $10.


seven


 

chris-o-five 12/5/2012 | 4:48:42 PM
re: Smart Is in for 2012

You don't have to replace every piece of equiptment, you just need to collect the data from those devices.  You then need to transform the data, so that for example, data from on manufacture's router looks like another manufactures data.  This way users get a consistent view and one that is meaningful to them because they have participate in defining the transformation requirements.


You also don't have to collect data from absolutely every piece of network gear, just the elements which have the biggest impact to customer experience.  


The reduction in capital comes from understanding where for instance you need to back haul data across your network and areas you don't, thus saving running fiber to every single cell tower.

StuBenington 12/5/2012 | 4:48:41 PM
re: Smart Is in for 2012




The majority of what exists today is infrastructure with stand-alone passive DPI probes; i.e. a “bump in the wire”.  This is valuable but operator feedback tells us that this is inadequate over the long term as services proliferate and become detached from specific providers.  A more forward-looking model incorporates this content awareness into the transmission network itself, either into edge routers or into network elements for mobile backhaul. 

This affords several benefits: first, this inspection can be throttled according to the needs of the operator, either for a sample of the traffic, on a per-flow basis (e.g. for stateful video sessions), or on a per-packet basis (e.g. for security gateway functions).  Next, incorporation into these elements provides a capex benefit (no additional elements), a scalability benefit (DPI scales directly with capacity), and an opex benefit (fewer nodes to manage, existing operational environment).  Finally, there are localization benefits: operators can respond directly and surgically to localized events affecting QoE, new revenues can be accrued through location mapping to service usage and time of day, and business models for networks like LTE can be augmented by associating content signatures with dedicated EPS bearers that assign QCIs based on specific performance parameters.


--Stu Benington, director, portfolio planning at Tellabs

HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE