Friday, 16 November 2007

VoIP: The Global appeal

According to a report entitled "Demand for VoIP increasing worldwide", Global revenues from IP telephony sales grew to $6,908 million (£3,383 million) in 2006. Western Europe formed the largest base of subscribers to these VoIP services, producing income for supplier amounting to $2,639 million (£1,292 million). Senior analyst of Point Topic John Bosnell suggested that the increase in revenues had been prompted by more customers taking up 'triple-play' service bundles which feature VoIP, broadband and TV in one package.

Asia Pacific was not far behind with USD 1.75 billion and North America with 2.41 billion revenue. Monthly ARPU was highest in North America at USD 20, followed by Western Europe and SE Asia with USD 15 each and the Asia-Pacific and Latin America with USD 10 per month each.
Dean Bubbley in his new report says that evolution of mobile VoIP will rapidly eclipse voice over WiFi and become a mainstream form of communication. The analyst firm predicts that the number of VoIPo3G users could grow from virtually zero in 2007 to over 250m by the end of 2012. This is comfortably in excess of the expected number of FMC users with dual-mode VoWLAN/cellular phones.

An interesting post in Forum Oxford by him highlights the following points:

First off, I'd better define what I mean - I see 4 manifestations of VoIPo3G:

- consumer or enterprise downloads to a 3G smartphone (or laptop) of software like Fring or Truphone or Yeigo, using 'over the top' VoIP on a flatrate data contract
- partner models (eg deals between Skype / Fring & a carrier or retailer or MVNO, or maybe some enterprise models if Cisco or Avaya or Microsoft does something) typically preloading a proprietary VoIP client on a handset or PC
- operator ‘non-telephony’ VoIP (eg PoC, multimedia telephony, some sort of other IMS voice app like conferencing)
- operator primary-telephony VoIP (eg for use on an all-IP network like LTE or UMB & to some extent HSPA+ or Rev A networks). * this is the really big one on a 5-year view but obviously dependent on 3.5G rollouts*

Basically LTE mandates VoIP - unless operators want to continue with a parallel circuit GSM or UMTS infrastructure in perpetuity. LTE-VoIP also the possibility of much higher spectrum efficiency for voice (in calls/MHz/cell) which is economically attractive given competition for spectrum & declining voice prices. It's also in the NGMN specs.
What I've tried to work out is what could happen in the time between cellular networks being capable of OK-quality VoIP in relatively small numbers (ie now) vs. the point when they're optimised for large-scale carrier grade mobile VoIP (ie quite soon on CDMA, quite a while on HSPA/HSPA+/LTE)

There's also issues like dependency on IMS, and whether operators want to simultaneously roll out a new radio tech (LTE) and transition their main revenue-earning service to VoIP - or if it makes sense to get some mobile VoIP experience before LTE rolls out. I think we'll see a lot more 3/skype type partnerships as 'toes in the water'.
I also reckon that an increasing number of 3G-enabled PC will be given operator VoIP client software for particular use cases (eg VoIP-enabled IM, or for other PC-based phone calls for conferencing etc)

Interesting to see
this Taiwan operator-led VoIPo3G service as an early example.
Probably worth saying as well that I'm not expecting all of an individual's telephony traffic to transition to VoIPo3G. Out of 3.5G coverage, it'll flip back to GSM or circuit UMTS for example, or might be WiFi indoors.

My view is that overall, VoIPo3G is more important than VoWLAN. The obligatory 'big number' headline is 250m users by end-2012 (ie about 7-8% penetration of global mobile subs and perhaps 20% of global 3G+ subs), mostly driven by the operators' own VoIP services.That's against my expectations of a low single-digit % for VoWLAN penetration. There's a lot more granularity but I'm not going to delve into it here now.
While VoWLAN has been qualitatively very important, the actual quantitative usage is really small. I don't see dual-mode phones or services being that prevalent for most operators. They're just too complicated to get right at a software level, because of the variety of ways in which WiFi is deployed (eg widespread use of private WiFi with its own security mechanisms). Whilst there's been a lot of work done over the last 3 years, most VoWLAN still has a lot of compromises.

I reckon that the same amount of work on VoIPo3G will yield more tangible results for both independent providers and the carriers themselves. It should also fit quite nicely with other trends like femtocells and mobile voice mashups (ie embedding VoIP in other applications). It can also be hidden "under the hood" so the user just sees a normal phone service - and therefore should be easier to integrate with things like prepay servers.

Not everyone agrees though. According to Russel Shaw, "At least not in North America, continent of my residence. Based on behavior patterns that significantly proceed VoIP, entrenched cell carriers will go to all marketing, regulatory, legal, lobbying and statutory costs to protect the current business models they believe are necessary to protect their investments in networks as they are now."

In other news, according to Yankee Group, the market for mobile internet services remains largely untapped, according to industry analysts which claim that operators and service providers could be raking in as much as $66bn per year, instead of the $9.5bn they are taking at present. Yankee Group warned that the full extent of the mobile internet services opportunity remains unexploited because service providers and their technology partners are not overcoming barriers to adoption quickly enough.

Finally, there is a warning for all VoIP users that the flexibility and openness ofSession Initiation Protocol (SIP) have made it a key building block forvoice-over-IP (VOIP) services, but SIP also makes carrier and enterpriseVOIP networks vulnerable to crippling attacks that could bring servicesdown for days, according to the latest report published by Light Reading'sVOIP Services Insider (, a subscriptionresearch service from CMP's Light Reading (

SIP is subject to the same types of attacks -- including viruses anddenial-of-service (DOS) attacks -- that affect email communications, but asuccessful attack through SIP is likely to have a larger impact on theaffected network, notes Denise Culver, research analyst with LightReading's VOIP Services Insider and author of the report. "SIP enablesvoice traffic to traverse VPNs, potentially carrying with it all of thethings a hacker might want to attach to such a message," she says.

"Whilethose in the email security world have had more than a decade to contendwith these issues, SIP security vendors are trying not only to address theissue of securing SIP messages but also to ensure that SIP can successfullytraverse a firewall at all."
A big part of the problem with SIP is that vendors have rushed productsinto the market that don't make use of all the security measuresrecommended in the protocol standard, Culver adds. The standard'sflexibility is also an issue in making networks vulnerable to securitybreaches, she says: "Until vendors reach a point at which interoperabilityis not just a requirement but actually something they recognize in terms ofthe security it provides across SIP itself, the protocol will remaininherently flawed."

No comments: