I've now had a chance to gather my thoughts about last week's Barcelona Bunfight. I can't say I enjoyed the trip this year (having only grudgingly decided to go right at the last minute), but it was nonetheless productive and informative. That said, it's still got to have been better in Barca, than if it was held in the other 2013-2017 MWC candidate cities of Paris, Munich or Milan.
For me, there were three standout themes:
- Handset OS's and ecosystems, especially the rise of Android, and the Nokia/Microsoft fall-out
- Ongoing discussion about the 3G/4G data traffic "problem", and how to add capacity, manage traffic and hopefully generate new revenue opportunities
- The future of personal communications, especially around VoIP on 3G/4G networks and the evolution of social networks and messaging.
Of course, there was a ton of other stuff going on as well - such as a lot of hype about mobile money / payments, the reality-distortion field around NFC, zillions of identiclone tablets, assorted fluffy apps/content things, and an absolutely bubble-tastic fleet of private jets at BCN airport.
Android gave everyone something to talk about & play with, while conveniently ignoring that the people with money (and especially discretionary technology income) still prefer Apple, while BlackBerry seems to be winning the hearts and minds of the next generation better than either. Nevertheless, the Android zone in Hall 8 was pretty much a full-on party, and maybe free smoothies can start to catalyse a sense of "cool" around the brand over time.
For me, perhaps the most interesting thing about Android isn't the high-end but the opposite. The sheer volume of designs and OEMs involved is driving the inevitable downward pressure on phone prices and margins. We can't be too far away from having a "good enough" smartphone at $100 price points - which then opens the market up to the vast prepay subscriber base. It also makes it harder for the operators to convince users to go for long contracts with subsidised devices - who's going to sell their soul for a 24-month contract (and a possibly locked-down device), just for the benefit of a $100 loan?
I'm not going to re-hash all the arguments about Microsoft / Nokia (apart from anything else, my input is going into something soon to be published by my friends over at Telco 2.0) but it was certainly on everyone's lips throughout the week. Overall, I'm cautiously positive, although the killing of Symbian seems to have been done a bit too abruptly. There's certainly going to be an uncomfortable transition period before any WP7 Nokia devices ramp up to take its place, plus the company seems to have lost an awful lot of developer goodwill. On the other hand, not dashing after the unproven non-Apple tablet space immediately seems very wise. My view is that there's going to be at least a year of disappointments around Android / WebOS / whatever tablets, and Nokia/Microsoft might be better off waiting to pick up the pieces in 2012 after the inevitable short-term bloodbath.
I stopped off to see HP and its new WebOS phones & tablet. Seem to be nicely engineered, although without *quite* the level of UI intuitiveness I might have expected, but I suspect that would change with practice. The larger phone with the QWERTY was quite appealing - the sort of thing I'd consider replacing my iPhone with next time, as it has a hint of individuality about it. In fact, I thought much the same about some of the WP7 devices as well.
(With impeccable timing, my iPhone black-screened and refused to restart, while I was standing at the Microsoft booth. Either the iPhone has a hidden "tantrum" feature triggered by detecting potential disloyalty from its owner, or else the guys from Redmond have special competitor-disruptor rays mounted on their stand).
One of my largest research themes in the last year has been around mobile data traffic management. It was at MWC in 2010 that I realised quite how many silo solutions there were - everyone had an answer to the "data tsunami" generated by dongles and smartphones, but there was no consistent approach to stitching together the various bits. Various flavours of offload, optimisation, policy, charging, connection managers, protocol tweaks and assorted other were around, but there was a lack of any "holistic" approach to blending these intelligently. I published a research note on this theme last May.
Since then, I've been continuing to see the evolution of the space, which has been fuelled by many operators' short-term needs for a "quick fix" to their data problems, coupled with some longer-term strategic thinking. I've been pretty vociferous about the claims of video optimisation as "the answer", in particular where it's done way back in the network without any decent awareness of the radio. I also have a lot of doubts about so-called "application-based policy", whereby operators can supposedly create "personalised" mobile data plans which include/exclude/prioritise specific traffic types or web destinations.
This trip to Barcelona enabled me to catch up with some of the latest developments around sub-topics such as offload and connection-manager software on devices, which I see as more strategic for overall traffic management than core-network heavy machinery. I was also struck by the fact that the largest believers in the "holistic" approach are actually the largest traditional network vendors, not some of the startups.
Ericsson's deal with Akamai looks truly important here, as in theory it should be able to combine video optimisation / caching with the ability to tweak policy right the way down to the scheduling algorithms in the base stations. All the boxes sitting in the core network or out on the Gi interface have a critical flaw - they neither understand the radio domain very well (eg is the user temporarily out of coverage, rather than in a congested area?), or else aren't in a position to "enforce" anything sufficiently granular to deal with the problem. Alcatel-Lucent is also tying various bits of its policy and content portfolio together - it bought CDN vendor Velocix last year, and also has some probe/DPI cleverness in the RAN.
More importantly, getting a firm like Akamai involved in video optimisation and policy is important, because unlike the operators, content publishers actually trust a CDN not to mangle or degrade video content without permission. Unnecessary transcoding or compression of video, without the awareness or permission of the producer, is extremely unpopular. While "in extremis" it may be necessary in times of congestion, it would still be better done with the involvement of the originator, not by the operator's network operating autonomously. I'm expecting to see some smarter content companies put notifications in their apps that monitor when telcos are "fiddling about" with their traffic, and inform the user of who is to blame if "artistic integrity" is compromised.
Among the policy & DPI vendors, I was particularly impressed by those that are offload-centric (eg Bridgewater), rather than those that are app-centric (eg Sandvine). I also saw some neat on-device software from the likes of Roke, which also optimises for cost/battery life as well as radio bearer availability.
It's time for an over-generalisation with a large kernel of truth here: network people don't *really* understand applications. They don't understand how users perceive applications, they don't understand how apps evolve and interact, they don't understand the limitations of their own boxes, and they don't understand the difference between an application and a traffic flow.
It's not their fault, to be honest - it's more the paucity of the English language for helping us distinguish between different grades of "things that run on top of platforms". Everything is an "application", from the silicon to the ecosystem as a whole. The mobile industry now has a layer-cake of different platform tiers and applications, and it's not the network vendors' fault that the app-stratum they can watch isn't as useful as the semantics would have them believe.
I lost count of the number of people suggesting that operators could have tiered services, say with one mobile broadband package optimised for "social networking". The descriptions of a theoretical "Facebook data plan" I heard from a few just don't stack up. For example, none of them had a good answer for me when asked what they'd be able to do now that Facebook forces its sites to encrypted SSL. Nor did they have a good answer about whether zero-rating Facebook traffic would include web links shared by friends, which are displayed *inside* the Facebook app on a smartphone. Or what would would happen if the June update of the FB app added something new, like video.
The other theme I was following at the show was the evolution of personal communications, and especially the role of operators and other participants. Various of my clients have been asking me to advise on "the future of voice" and new business models which take account of fully IP-ised networks like LTE. Much of the discussion is around new voice apps and revenue streams, and especially the growing distinction between "voice" in general and "telephony" as merely one specific voice application.
As expected, we heard a fair amount of noise about VoLTE for voice on LTE, as well as the news that T-Mobile was dropping its cherished advocacy of VoLGA. There were also quiet a few companies quietly pitching non-IMS approaches to LTE VoIP, which I think have a strong chance of adoption. I'm extremely skeptical about the suggested timelines I've heard from VoLTE, which seem to be driven by the needs of both PR and the desire to foster consensus through inertia (or at least collectivising the risk of failure). In my view, the likelihood of getting normal handsets into the market, running VoLTE for "primary mobile telephony" with good quality, battery life & mobility, before 2013 is very slim indeed. Yes, even on Verizon. Nevertheless, I do agree that VoLTE will eventually happen, for some operators in some contexts - although it will certainly never be ubiquitous. My post-MWC views are essentially unchanged since the pre-MWC post & lengthy discussion in the comments here.
If VoLTE doesn't deliver, I wouldn't be surprised if VoLGA gets reincarnated at some point - probably with a face-saving rebrand. I'm sure T-Mobile knows this as well. (I tried to come up with an amusing pun for a future DaNUBE or THaMeS acronym for LTE Voice, but I haven't managed it yet - suggestions welcome).
On the messaging and social-network side, MWC included a lot of discussion about integrating incumbent platforms such as Skype and Facebook, as well new telco-centric niche efforts such as the GSMA's RCS-e (a new revised version). My views on RCS have been pretty consistent since its announcement 3 years ago, so it's good to see that finally the GSMA has ditched the focus on presence because it kills the phone battery and floods the network with signalling. I'm still wading through the new specs, but despite the high-profile announcement of future operator support and a few demos, I still think it is too little, too late.
And there's still no involvement from the key players in messaging and social communications such as Facebook, Apple, RIM or Skype - the operators need to stop the ridiculous "them and us" stance and prove their credentials in social communications, interoperating via the web, and gaining viral adoption because users find the services valuable. For RCS to succeed there also need to be a firm commitment to a "freemium" business model, and a route to getting away from the legacy of the "phonebook" - a hundred-year old way of viewing your social affiliations, that is rapidly becoming obsolete. Anyone wanting a full critique of RCS should obtain a copy of my report from the end of last year, available here.
For me, there were three standout themes:
- Handset OS's and ecosystems, especially the rise of Android, and the Nokia/Microsoft fall-out
- Ongoing discussion about the 3G/4G data traffic "problem", and how to add capacity, manage traffic and hopefully generate new revenue opportunities
- The future of personal communications, especially around VoIP on 3G/4G networks and the evolution of social networks and messaging.
Of course, there was a ton of other stuff going on as well - such as a lot of hype about mobile money / payments, the reality-distortion field around NFC, zillions of identiclone tablets, assorted fluffy apps/content things, and an absolutely bubble-tastic fleet of private jets at BCN airport.
Android gave everyone something to talk about & play with, while conveniently ignoring that the people with money (and especially discretionary technology income) still prefer Apple, while BlackBerry seems to be winning the hearts and minds of the next generation better than either. Nevertheless, the Android zone in Hall 8 was pretty much a full-on party, and maybe free smoothies can start to catalyse a sense of "cool" around the brand over time.
For me, perhaps the most interesting thing about Android isn't the high-end but the opposite. The sheer volume of designs and OEMs involved is driving the inevitable downward pressure on phone prices and margins. We can't be too far away from having a "good enough" smartphone at $100 price points - which then opens the market up to the vast prepay subscriber base. It also makes it harder for the operators to convince users to go for long contracts with subsidised devices - who's going to sell their soul for a 24-month contract (and a possibly locked-down device), just for the benefit of a $100 loan?
I'm not going to re-hash all the arguments about Microsoft / Nokia (apart from anything else, my input is going into something soon to be published by my friends over at Telco 2.0) but it was certainly on everyone's lips throughout the week. Overall, I'm cautiously positive, although the killing of Symbian seems to have been done a bit too abruptly. There's certainly going to be an uncomfortable transition period before any WP7 Nokia devices ramp up to take its place, plus the company seems to have lost an awful lot of developer goodwill. On the other hand, not dashing after the unproven non-Apple tablet space immediately seems very wise. My view is that there's going to be at least a year of disappointments around Android / WebOS / whatever tablets, and Nokia/Microsoft might be better off waiting to pick up the pieces in 2012 after the inevitable short-term bloodbath.
I stopped off to see HP and its new WebOS phones & tablet. Seem to be nicely engineered, although without *quite* the level of UI intuitiveness I might have expected, but I suspect that would change with practice. The larger phone with the QWERTY was quite appealing - the sort of thing I'd consider replacing my iPhone with next time, as it has a hint of individuality about it. In fact, I thought much the same about some of the WP7 devices as well.
(With impeccable timing, my iPhone black-screened and refused to restart, while I was standing at the Microsoft booth. Either the iPhone has a hidden "tantrum" feature triggered by detecting potential disloyalty from its owner, or else the guys from Redmond have special competitor-disruptor rays mounted on their stand).
One of my largest research themes in the last year has been around mobile data traffic management. It was at MWC in 2010 that I realised quite how many silo solutions there were - everyone had an answer to the "data tsunami" generated by dongles and smartphones, but there was no consistent approach to stitching together the various bits. Various flavours of offload, optimisation, policy, charging, connection managers, protocol tweaks and assorted other were around, but there was a lack of any "holistic" approach to blending these intelligently. I published a research note on this theme last May.
Since then, I've been continuing to see the evolution of the space, which has been fuelled by many operators' short-term needs for a "quick fix" to their data problems, coupled with some longer-term strategic thinking. I've been pretty vociferous about the claims of video optimisation as "the answer", in particular where it's done way back in the network without any decent awareness of the radio. I also have a lot of doubts about so-called "application-based policy", whereby operators can supposedly create "personalised" mobile data plans which include/exclude/prioritise specific traffic types or web destinations.
This trip to Barcelona enabled me to catch up with some of the latest developments around sub-topics such as offload and connection-manager software on devices, which I see as more strategic for overall traffic management than core-network heavy machinery. I was also struck by the fact that the largest believers in the "holistic" approach are actually the largest traditional network vendors, not some of the startups.
Ericsson's deal with Akamai looks truly important here, as in theory it should be able to combine video optimisation / caching with the ability to tweak policy right the way down to the scheduling algorithms in the base stations. All the boxes sitting in the core network or out on the Gi interface have a critical flaw - they neither understand the radio domain very well (eg is the user temporarily out of coverage, rather than in a congested area?), or else aren't in a position to "enforce" anything sufficiently granular to deal with the problem. Alcatel-Lucent is also tying various bits of its policy and content portfolio together - it bought CDN vendor Velocix last year, and also has some probe/DPI cleverness in the RAN.
More importantly, getting a firm like Akamai involved in video optimisation and policy is important, because unlike the operators, content publishers actually trust a CDN not to mangle or degrade video content without permission. Unnecessary transcoding or compression of video, without the awareness or permission of the producer, is extremely unpopular. While "in extremis" it may be necessary in times of congestion, it would still be better done with the involvement of the originator, not by the operator's network operating autonomously. I'm expecting to see some smarter content companies put notifications in their apps that monitor when telcos are "fiddling about" with their traffic, and inform the user of who is to blame if "artistic integrity" is compromised.
Among the policy & DPI vendors, I was particularly impressed by those that are offload-centric (eg Bridgewater), rather than those that are app-centric (eg Sandvine). I also saw some neat on-device software from the likes of Roke, which also optimises for cost/battery life as well as radio bearer availability.
It's time for an over-generalisation with a large kernel of truth here: network people don't *really* understand applications. They don't understand how users perceive applications, they don't understand how apps evolve and interact, they don't understand the limitations of their own boxes, and they don't understand the difference between an application and a traffic flow.
It's not their fault, to be honest - it's more the paucity of the English language for helping us distinguish between different grades of "things that run on top of platforms". Everything is an "application", from the silicon to the ecosystem as a whole. The mobile industry now has a layer-cake of different platform tiers and applications, and it's not the network vendors' fault that the app-stratum they can watch isn't as useful as the semantics would have them believe.
I lost count of the number of people suggesting that operators could have tiered services, say with one mobile broadband package optimised for "social networking". The descriptions of a theoretical "Facebook data plan" I heard from a few just don't stack up. For example, none of them had a good answer for me when asked what they'd be able to do now that Facebook forces its sites to encrypted SSL. Nor did they have a good answer about whether zero-rating Facebook traffic would include web links shared by friends, which are displayed *inside* the Facebook app on a smartphone. Or what would would happen if the June update of the FB app added something new, like video.
The other theme I was following at the show was the evolution of personal communications, and especially the role of operators and other participants. Various of my clients have been asking me to advise on "the future of voice" and new business models which take account of fully IP-ised networks like LTE. Much of the discussion is around new voice apps and revenue streams, and especially the growing distinction between "voice" in general and "telephony" as merely one specific voice application.
As expected, we heard a fair amount of noise about VoLTE for voice on LTE, as well as the news that T-Mobile was dropping its cherished advocacy of VoLGA. There were also quiet a few companies quietly pitching non-IMS approaches to LTE VoIP, which I think have a strong chance of adoption. I'm extremely skeptical about the suggested timelines I've heard from VoLTE, which seem to be driven by the needs of both PR and the desire to foster consensus through inertia (or at least collectivising the risk of failure). In my view, the likelihood of getting normal handsets into the market, running VoLTE for "primary mobile telephony" with good quality, battery life & mobility, before 2013 is very slim indeed. Yes, even on Verizon. Nevertheless, I do agree that VoLTE will eventually happen, for some operators in some contexts - although it will certainly never be ubiquitous. My post-MWC views are essentially unchanged since the pre-MWC post & lengthy discussion in the comments here.
If VoLTE doesn't deliver, I wouldn't be surprised if VoLGA gets reincarnated at some point - probably with a face-saving rebrand. I'm sure T-Mobile knows this as well. (I tried to come up with an amusing pun for a future DaNUBE or THaMeS acronym for LTE Voice, but I haven't managed it yet - suggestions welcome).
On the messaging and social-network side, MWC included a lot of discussion about integrating incumbent platforms such as Skype and Facebook, as well new telco-centric niche efforts such as the GSMA's RCS-e (a new revised version). My views on RCS have been pretty consistent since its announcement 3 years ago, so it's good to see that finally the GSMA has ditched the focus on presence because it kills the phone battery and floods the network with signalling. I'm still wading through the new specs, but despite the high-profile announcement of future operator support and a few demos, I still think it is too little, too late.
And there's still no involvement from the key players in messaging and social communications such as Facebook, Apple, RIM or Skype - the operators need to stop the ridiculous "them and us" stance and prove their credentials in social communications, interoperating via the web, and gaining viral adoption because users find the services valuable. For RCS to succeed there also need to be a firm commitment to a "freemium" business model, and a route to getting away from the legacy of the "phonebook" - a hundred-year old way of viewing your social affiliations, that is rapidly becoming obsolete. Anyone wanting a full critique of RCS should obtain a copy of my report from the end of last year, available here.
What about RhINE: Really we'll have IMS Networks Eventually
ReplyDeleteDean B. has earlier questioned the use of Cisco numbers to predict the growth of data. Analysys Mason late last year published data based on operator's statistics showing a more modest growth in data.
ReplyDeleteSeveral operators e.g. Elisa in Finland have verified this. They see a linear growth at the moment. The reason due to Elisa is that the main growth comes from new customers and they don't attract too many new dongle-customers. Smartphones consume less data.
This is a discussion that I miss. Not the least in BCN.
(I agree with all that Dean is writing about BCN).
Disclaimer: I work for Ruckus Wireless, but I figure I'll do a shameless plug.
ReplyDeleteInterested that in your approach to Mobile Data 3G/4G Offload you missed Ruckus Wireless's Stand. I am not surprised you missed a small vendor like Ruckus but rather that you missed the company who won Best Mobile Broadband Technology at the conference, and whose solution was quite honestly enough to give most vendors who walked through it a pretty big hard on. You could say it even provided the 'holistic' approach you mentioned was so lacking... check it out :-)
*hits forehead*
ReplyDeleteNailing IMS to LTE = NILe
How could I have missed the obvious dead parrot reference?
Kent - interesting point. I'd also expect the growth pattern to be non-linear. As "laggards" adopt mobile broadband, they will tend to drag the averages down compared to the early-adopter centric market. Then over time, as all users & devices evolve, you expect to see a steady increase again. Sort of a sideways-S shape on the chart.
Anonymous - I did meet with Ruckus & had a very interesting discussion with them. I agree with some of what they say, but I have significant doubts about I-WLAN style approaches to backhauling WiFi traffic through the operator core. I'll be writing a post about WiFi offload soon.
In general, I have always thought that WiFi is much more important (and a better fit) for fixed operators than mobile, and that remains the case.
Dean