iPhone Troubles Might Go Beyond Signaling

What's ailing the AT&T Inc. (NYSE: T) 3G network might not be fixed with the simple backhaul upgrade that the carrier is touting. But it's possible that boosting the signaling capacity, which has been flagged as an alternative answer, won't do the trick either.

It depends on whom you talk to, of course. Nokia Networks claims it's got the surprisingly simple answer for the signaling problem, but Alcatel-Lucent (NYSE: ALU) tells Light Reading that AT&T's problems are probably multifaceted and can't be fixed in one blow.

The issue at hand is the rate of dropped calls and bad connections on the 3G network, a problem that's caught fire since Apple Inc. (Nasdaq: AAPL) released the 3G iPhone. (See AT&T Mobile Boss: NYC & San Fran Are 'Underperforming' and AT&T to Spend $2B More on Wireless in 2010.)

And now, Apple has released the iPad, at least in the United States, leading to concerns that things might get worse. (See Apple to Ship New 3G iPads by May 7, Will the Apple iPad Crush 3G Networks?, and Apple Delays International iPad.)

Ars Technica mentioned the signaling factor in February. The iPhone frequently disconnects its data connection in order to prolong battery life. But that means that to do tasks like check for new emails, the device has to reintroduce itself to the network, restarting the secret handshake that requires lots of signals traded back and forth. The result is the equivalent of thousands of new phone calls, and it could be overloading the network's signaling capacity. (See What if Capacity Isn't AT&T's iPhone Problem?)

Folks at the 3rd Generation Partnership Project (3GPP) predicted this problem, it turns out, and there's a standardized, years-old fix for it, according to Phil Twist, Nokia Siemens's head of network systems marketing and communications.

It's called the paging channel, or Cell PCH, and it's been in Nokia Siemens gear for five years, with more than 50 operators using it, Twist says. It's just a simpler standby mode, one that uses fewer signaling calls to ping the network.

"It's a simple software feature that reduces the overall congestion of the network by a factor of three and doubles the user's battery life," Twist says. "It's been in every handset shipped since 2007."

But as Twist noted in this message board post, and as he explained to Light Reading at the time, equipment vendors haven't bothered to implement it. Equipment vendors other than Nokia Siemens, that is. Instant publicity blast!

So, it would appear the signaling problem will remain unaddressed unless AT&T's radio network controller (RNC) vendors -- Alcatel-Lucent and Ericsson AB (Nasdaq: ERIC), according to Heavy Reading analyst Gabriel Brown -- can provide the software upgrade to incorporate the paging channel. (This might be a convenient point to mention that AT&T declines to comment on the whole signaling issue.)

But wait. AlcaLu is quick to point out that signaling isn't likely to be the only thing affecting calls on AT&T's network.

A rainbow of troubles
Networks are as different as fingerprints, and there isn't always one specific cause of bad service, says Mike Schabel, an Alcatel-Lucent general manager.

While he won't talk about AT&T specifically, Schabel says there's a long list of problems he's seen in general: security problems introduced by laptops, wireline applications that prove inefficient on a wireless network, or plain old misconfigurations. "There's never been, in any network I've gone to, the same root cause," he says.

As you'd expect, AlcaLu claims it's got just the tool for the trick: the 9900 Wireless Network Guardian. It's a software package that runs on "semi-custom" hardware at one location in the network, Schabel says. From there, it measures each network element's contribution to bad service, letting operators decide where to prioritize their corrections.

Schabel works on test gear like the 9900, not wireless gear, so he doesn't have a comment about whether AlcaLu will be implementing Cell PCH. But he does have an opinion: He thinks Cell PCH, in any network, is one small fix applied against a group of big problems -- not a cure-all, in other words.

So, maybe AT&T does need to add more capacity and more backhaul to its 3G network, and maybe it does need to cool down the signaling chatter among smartphones. But if Schabel's world view is right, neither fix is going to put the iPhone Effect to rest.

— Craig Matsumoto, West Coast Editor, Light Reading

Page 1 / 2   >   >>
DCITDave 12/5/2012 | 4:38:58 PM
re: iPhone Troubles Might Go Beyond Signaling

You sure it's not Captain Phil Twist?

Could be a big player in Project Cadmus (Justice League). I'm just sayin'.

Pete Baldwin 12/5/2012 | 4:38:54 PM
re: iPhone Troubles Might Go Beyond Signaling

On a more serious note,

I still don't have a complete handle on what makes the network perform badly.  But I do like AlcaLu's explanation.  Figuratively speaking, it's an analog problem -- lots of tweaks and fine-tuning required, and no single switch that can be flicked from zero to 1 to make it all better.  That would make a lot of sense.

Anyone else have thoughts? How hard is this going to be for AT&T to fix?

Pete Baldwin 12/5/2012 | 4:38:54 PM
re: iPhone Troubles Might Go Beyond Signaling

Geeky as I might be, I had to look up Cadmus.  Point for Phil (in the race to the depths of nerddom...)

Michelle Donegan 12/5/2012 | 4:38:53 PM
re: iPhone Troubles Might Go Beyond Signaling

AlcaLu has some recent experience with what can go seriously wrong on a 3G network after the fiasco in New Zealand …


But that was more than just poor network performance. Anyone know the latest on that situation, btw?

Michelle Donegan 12/5/2012 | 4:38:52 PM
re: iPhone Troubles Might Go Beyond Signaling Presumably all the extra signalling chatter on the network is a problem for all operators. Guess I'm surprised more operators talking about it.
cebert 12/5/2012 | 4:38:51 PM
re: iPhone Troubles Might Go Beyond Signaling Overall the iPhone has done our industry a great favor in a way not typically discussed. Simply, we must flip our telescopes and not focus so closely on mobile network usage trends, but rather the broader Internet. The point highlighted by Phill is one item in a broader understanding. So ALU's point around this being a multifaceted issue is complimentary. The bottom line is supporting a 'single internet' regardless of bearer means the right understanding of mobility and IP (not MobileIP) and the inherent responsibility of both network vendors and carrirers to have a 'network of one' vision that is pervasive in the architecture and doesn't gloss over the differences and required to make it all transparent to the end user, ultimately. (sent from iPad - can't seem to review so pardon in advance and case in point for our friends above the network layer)
phil-tw 12/5/2012 | 4:38:50 PM
re: iPhone Troubles Might Go Beyond Signaling

Happy to see I've apparently been nominated for promotion to Captain :-)

Seriously, let me just pick up on the points in the article. By using the cell_PCH function standardised by 3GPP it is entirely possible to make the signaling load on networks heavy with smart phones reduce to the level where it is virtually no longer a consideration. And by reducing the signaling load you also reduce the background battery consumption on the smart phone - we've seen it DOUBLED in real life situations (and yes: that's doubled comparing one on our network with cell_PCH, with that of a certain French-US competitor's network only just built). Battery life is the #1 user concern if your smart device can't make it through the day. Second only to not being able to use it at all because the network is busy.

Of course it's clever to have a way to identify which components in the network are close to being overloaded - and naturally at Nokia Siemens Networks, we have a whole set of smart ways to do that as well. But isn't it far better to avoid the unnecessary load in the first place? Imagine you are sitting on Route 101 unable to get to your office because of the traffic. We have a way to take two thirds of the cars off the highway: instant fix, no disruption. What's discussed in this article is a way to show you where you could build extra lanes to make the highway wider….

And - yes of course networks can get congested by packet data for many reasons, and there are many solutions eg:

 - because you have filled up the available radio capacity on a base station (so you need to switch on another carrier, add sectors or add base stations)

 - because the link from the base station to the RNC is full (so add more E1/T1 links, or increase the microwave capacity, or switch on another colour on your optical link)

 - because the RNC reaches one of its design limits - throughput, signalling, … (so bypass the RNC for data traffic with Nokia Siemens Networks' natty i-HSPA solution or of course switch on cell_PCH if you haven't already. Or shape the traffic with some intelligence in the servers. And only as a last resort add more RNCs)

 - because the core network reaches one of its design limits - throughput, transactions per second, number of simultaneous sessions (so use NSN's Flexi-NG packet core solution which not only leads the way in those 3 dimensions but also in the fourth: traffic prioritisation and QoS controls so the real-time traffic or the 'gold service' customers go through ahead of the non-real time traffic which can tolerate a short delay).

The rapid growth in the number of smart devices and the number of applications they support is a great opportunity - hence we are running ahead of the wave and designing networks components and features to take advantage of this: allowing the users to make the most of their devices and the service providers to align their investments to growth in revenue.

Phil Twist - Nokia Siemens Networks

Pete Baldwin 12/5/2012 | 4:38:50 PM
re: iPhone Troubles Might Go Beyond Signaling

I'd agree with your 'network of one' perspective, cebert.

Equipment vendors love the 'network of one' idea, too, but tend to twist it into a 'network of One Vendor' pitch.  (Cisco.)  How could the industry get around that?  Probably requires carriers to get even more involved.

This all reminds me (yet again) that i should get back in touch with the IPsphere Forum folks.

freetoair 12/5/2012 | 4:38:44 PM
re: iPhone Troubles Might Go Beyond Signaling

what? is they a cipher key needed to understand that post?

Juipter to Mars - can you hear me?

Pete Baldwin 12/5/2012 | 4:38:43 PM
re: iPhone Troubles Might Go Beyond Signaling

Phil - thanks for the posting.  I agree that it's best to avoid the strain on the network in the first place, but I wonder if wireless data is going to simply be a capricious beast for some time, hampered by lots of separate problems that add up.

Page 1 / 2   >   >>
Sign In