Pages

Pages

Thursday, August 27, 2015

The Internet is now attacking those that try to damage it, not just avoiding them

It used to be said that "The Net interprets censorship as damage and routes around it" - a quote by John Gilmore, in 1993.

In my view, the Internet community is now maturing to a state where it is going on the offensive rather than just taking defensive avoidance strategies. 

When it is threatened, rather than just looking for ways to mitigate or "route around" potential damage, it is now going after the source of that damage (& its promoters). It is going beyond mere nullification to actually escalating the battle - either codifying methods pre-empting further problems from the same source, or actually attacking that group more broadly.

A prime example has been the rise of ad-blocking. Most Internet users are OK with the general principle of passive, neutral, non-invasive advertising. It's a bit annoying, but no more so than ads in public spaces like poster hoardings. But what gets resented is where those ads are privacy-invasive (eg with tracking), hugely interruptive (pop-ups) or consume a lot of resources (eg auto-playing video).

As a result, the Internet has given birth to a variety of tools which don't just block the annoyances, but take out the basic, more-acceptable formats too, as collateral damage. The Internet ad industry has over-reached the boundaries towards a form of pseudo-censorship (or at least, "mucking around with content" in undesirable ways), and found its fundamentals under attack. The idea that telcos will charge advertisers for traffic, on threat of blocking, is another example of something that will backfire even more strongly, as I wrote recently.

Another area is telcos'/ISPs' attempts to "monetise" Internet connections, or "optimise" traffic, in wholly self-serving fashion. The use of DPI for filtering, advert insertion with HTTP man-in-the-middle proxies, and various other unwanted forms of interference, has led to the wholly-predictable rise of encryption everywhere. We see HTTP2 with Google's SPDY baked-in, wider use of VPNs and so forth. The irony here is that this also has collateral damage, preventing some forms of actually-useful forms of network management. Again, strategic over-reach has resulted in the Internet not just defending its basic utility against a form of censorship, but escalating the battle.

What prompted this post was reading a new IETF Draft (see here) which explicitly "mandates end-users as the highest priority constituency for Internet standards". The needs of users (who generally aren't well-represented / advocated in standards) are now being absolutely prioritised above ISPs, vendors, developers, "implementors", governments and so forth. Theoretical considerations of technical "purity" or "elegance" come right at the bottom of the list.

It means that new IETF standards ought to check - and explicitly state - that user welfare is not harmed at the expense of other groups' gain. It's a bid to stop some of the back-room chicanery that goes on inside a lot of standards development (many groups are a lot worse than IETF - I'm sure readers will be able to identify the real culprits here). Small technical details are sometimes deliberately made to capture value on behalf of other constituencies, to the detriment of users (or rival groups). 

This forces the principle of "design thinking" onto the normal engineering-led (and perhaps commercially-influenced) process.

This move appears to have been catalysed, in part, by the ham-fisted opposition to HTTP2/SPDY by the ironically-named "Open Web Alliance" set up by various (mostly US) telcos and vendors as part of ATIS. This group came up with the risible notion of "trusted proxies" which did not just allow useful traffic-management, but started with the premise of allowing ISPs free rein to break encrypted streams for a variety of use-cases, including ad-insertion, big-data collection, application discrimination and various other questionable practices.

In other words, the Internet is not just neutralising a specific example of "damage", but it's trying to ensure that those that attempted to damage it are de-fanged for the future as well. 

The US battle against Net Neutrality was another instance. An attack by the ISP/telco industry on a relatively straightforward legal rule ended up with a much larger fight-back: culminating in the FCC implementing Title II for broadband, with help along the way from the likes of John Oliver and a major grass-roots campaign. What could have been a minor defeat turned into a major strategic rout, because of push-back to the way the US regulatory and lobbying system works. Whether it has learned its lesson is yet to be seen.

Various repressive regimes have also learned to their cost that attempting to harm the Internet can result in disproportionate retribution from it in response.

Over-doing the use-cases for zero-rating has also attracted the Internet community's ire, meaning that valid use-cases may suffer along with questionable ones, as some regulators ban the concept entirely.

Then there's the spies. Bodies like the NSA & GCHQ have had to deal with far more, and stronger, crypto, because recalcitrant vendors, whose reputations and exports have been hurt by intercepted data or underhand hacks, responded to users' demands uncompromisingly. The intelligence agencies - who have a difficult and thankless task - committed the cardinal sin of over-reach, and then had to deal with the consequences.

It seems likely that some of the push for LTE-U (4G in the same spectrum as WiFi) may go the same way. By angling for a version called LAA which needs licenced spectrum as a pre-requisite, the industry may well find itself unable to use LTE-U at all once the regulator gets involved. Plus it might even get worse - by suggesting WiFi and LTE can "play nicely" together (with coexistence protocols) in unlicenced bands, perhaps we'll get escalation. Surely this means private WiFi could run in today's licenced bands as well, as the evidence shows it could be implemented in ways that wouldn't interfere? That could hypothetically aid overall efficiency of spectrum utilisation, whilst not damaging the value of carriers' own spectrum.... 

The strategy of arguing "Who, me? Innocent little me? My intentions are good, honest!", with fingers crossed behind ones back, is no longer a viable one.

It could be argued that with these new powers, the Internet is in danger of becoming more than just petulant, but risks sometimes throwing the baby out with the bathwater. That seems to be the case with some of the genuine security risks that might emerge as an unintended consequence of crypto. It's possible that the Internet will need to control its own power better, a bit like the Incredible Hulk. 

But for now, until it gets that self-control, the lawyers and lobbysists, telcos and ISPs, advertisers and snoops all need to imagine the Internet saying "You wouldn't like me when I'm angry" before pushing their luck too far. If you "try it on" with the Internet, prepare for a bigger slap than you expected. So for example, all this nonsense about "level playing fields" (see here), erecting toll-gates (here), or sly attempts to influence lawmakers about spurious notions of "platform neutrality" (here) will make the Internet bite back. Hard.

Monday, August 24, 2015

Deutsche Telekom's disappointing visions for telephony, VoLTE & WebRTC. And it still backs RCS

I've just listened to a webinar by RCRWireless, covering opportunities for VoLTE, RCS, WebRTC & WiFi-calling. The link is here. It's an hour long and worth a listen.

As well as the RCR editor Dan Meyer, it has comments from another analyst (Mike Thelander from Signals), a vendor (Comverse), it also included participation from Deutsche Telekom's head of voice and messaging - who was also the former head of RCS at the GSMA.

There's a lot of angles I could cover in discussing the content of the call, some which I agree with, but also quite a lot that I don't think is right.

The positives:

 - Nobody really seems to believe that ViLTE, aka IR.94 IMS-based video-calling, is a winner in its current form.
- VoLTE has some benefits as a basis of LTE telephony, notably faster call setup time than circuit fallback, and the potential to (eventually) allow operators to refarm spectrum.
- However, VoLTE is still difficult to implement, and even harder to interoperate between telcos because there's different flavours of it. I agree that interop will help
- NFV is helpful in reducing costs of IMS & VoLTE. I'd agree with this - cloud implementations will generally be "least-worst" and can also scale up with demand, which should help limit the downside from "peak telephony" impacting investment ROI
- DT recognising that some communications is "primary" and some is "secondary". Yes, but see below. 

 The so-so's:

- VoWiFi is useful. Yes, it is, but that's hardly news. It's been around in various guises for 10+ years, and simply combining it with VoLTE for better indoor coverage isn't a big deal. The sudden hype around 3GPP WiFi calling deserves a whole extra blog post, as it ties in with the concept of primary / secondary comms discussed below. Extra revenue? No.
- WebRTC will help with service extension and also the creation of new comms services via APIs. Yes absolutely, but those are only some of WebRTC's use-cases for telcos. The main opportunities arise from new areas outside traditional telephony models. Look at Comcast's video-streaming, Telenor's appear.in, Telstra's telemedicine apps or Telefonica's TokBox platform as examples

- VoLTE is a "must". Well, for some operators, especially if they're either really spectrum-constrained, or they lack 2G/3G coverage everywhere they've got LTE, so can't use fallback. But for most, it's still a distress purchase that costs money, without bring any new revenues or real customer benefits to the table. If spectrum, regulations and licence terms allow, operators should either just stick to 900MHz GSM in perpetuity for "plain old telephony", or perhaps take a major leap to either "bring your own voice" models,  or outsource it to a 3rd-party cloud-voice provider, as some fixed/cable players are doing.
- DT mentioned something it's working on called "enhanced voice", with extra features pre/post call and the ability to drop content into a voice interaction mid-app. Apparently they're working with Orange and device vendors like Sony & Samsung. It apparently combines RCS with VoLTE - and there's some hints it might allow some more granularity and hints of context, such as allowing an "urgent" flag to be added pre-call. I'll reserve judgement until I see it (allegedly towards the end of 2015)
- Oblique references / hopes to Apple becoming more supportive of IMS (especially its VoLTE implementation), with what sounds like a relatively forlorn hope that it might put native RCS into iPhones - but probably only if & when RCS gets enough support from users to make it worth the effort. As I've said before, RCS needs to earn its ubiquity - it can't just assume it will become ubiquitous because the telcos say so.

The negatives:

- The continued obsolete framing of "telcos vs. OTTs". This narrative is dead. The only argument advanced for why anyone might switch back from alternative voice or messaging to telco-offered services was that they don't incur data charges. Given that 50-90% of use is on WiFi anyway, that's not very convincing. Very little discussion about functionality, UX, or user design.
- No argument why operators can survive slowly on "primary" communications, given that current evidence show that secondary alternatives are becoming much more important, and WebRTC, contextual comms and various 3rd-party apps are making them ever better. Worse, there was no analysis of what % of current voice or messaging traffic is actually "primary use-cases", and what is left exposed to more-functional (but less-ubiquitous) competing alternatives. My view is that we're probably let with 20-40% of historic voice traffic, and maybe 10% of SMS's, once all the "secondary" uses have been siphoned off. In some cases, secondary comms will be more valuable than primary - it's wrong and misleading to use semantics that imply value and importance.
- No distinction was made between "voice" as a broad media type, and the specific model of "calling" we are familiar with in 130yr old telephony. No reference was made to contextual communications, new interaction modes, hypervoice, ephemeral communications - or even conferencing, for that matter. All the speakers seemed stuck in a voice=call mentality.
- There was hardly any mention of developers, although Comverse talked about APIs and revenue-shares. (It also mentioned sponsored data, more negatively)
- DT was quite negative about WebRTC, advancing a straw-man argument that it's not appropriate for primary communications. Well, it's not a direct replacement for primary telephony on mobile phones yet, I'll agree, but it's certainly fine in enterprise situations like call-centres, numerous innovative mobile apps (especially video-based ones) and more importantly, it's where the innovation & disruption is. DT has been working on various prototypes using WebRTC & discussing them at conferences for ages now - it needs to get them out there, perhaps focusing on business units outside the reach of the conservative IMS dinosaurs. Learn by doing, and perhaps find some new niches in enterprise or TV or elsewhere. Look at NTT, Telenor, Telstra, Telefonica, Comcast & others who are being aggressive rather than taking a narrow, historical view of telcos' role in communications apps
- RCS. Unsurprisingly, the ex-head of RCS at GSMA is still enthusiastic about it at DT. It's still dead, despite the attempts by speakers to assert a "resurgence" of interest. Yes, there's a predictable attempt to bolt it onto the slow & grinding uptick in VoLTE, but no, there's no reason why anyone would ever use it. Pretending otherwise is fooling nobody - a supposed 270% increase in active use from what is almost certainly a pathetically-low base is an irrelevant figure. Unless hard numbers on MAUs/DAUs are issued about genuine RCS use (not just IP-messaging infrastructure as SMSC replacement) then it's safe to assume it's a continued failure.

- As ViLTE/IR.94 is lacking in practicality, the focus of video-calling will be around RCS instead. I'm running out of palms to slap my face on this.... (Hint: call Telenor and licence/rebrand appear.in instead, and save yourself 5 years of pain, millions of euros, and maybe get some upside at the end of it)

Overall, the webinar had some interesting things - but also revealed the lack of ambition and vision among service-provider folk when it comes to exploiting and leading the "future of communications". 

I see where Signals is coming from about spectrum refarming and capacity efficiency for VoLTE vs. so-called "OTT apps", but if makes a false equivalence and assumes they offer the same product/purpose for the end-user. They don't, in general. People pick the best/cheapest tool for the job. Nobody will switch from SnapChat video or Talko voice to a less-featured, less-cool telco alternative based on VoLTE, even if it has QoS and is zero-rated. Having guaranteed "quality" for the wrong service doesn't improve experience. It just gives a fallback option if nothing better is available. Most communications won't look like a "call" at all.

Comverse appears to be being led by the path of least resistance from some customers (some of whom presumably are asking for RCS) rather than making a stand and propounding a bold vision about what's really possible with the future of communications. The talk about "fourth wave" digital services with connected cars & healthcare didn't articulate why those opportunities should fall to IMS-based telcos rather than external innovators designing for the problem, not the platform. Hopefully the Acision acquisition - including the more developer/enterprise-focused Forge WebRTC platform - will nudge them in the right direction, rather than vice versa. More focus on context, or vertical applications unencumbered by legacy telco mindsets, will help.

The DT position was very much about standardised, interoperable incrementalism, trying to eke out 130 years of telephony and 20 years of SMS into mildly-updated forms. While there's definitely going to be a long tail for "vanilla" "primary" voice calls, I expect the revenue to plummet as supply massively exceeds demand. Meanwhile, denigrating new use-cases for communications as "secondary" pretty much guarantees exclusion from the vanguard any new and exciting opportunities. Disappointing, but unsurprising.

EDIT: A comment on the LinkedIn version of this post made me think about how DT's strategy might play out. Hopefully it will be relatively easy for wiser & more visionary units of DT to act autonomously, and ignore the HQ dinosaur herd. Telefonica has managed it, from time to time, with groups like Tuenti and TokBox following their own paths. Similar story for Orange Business Services or the Vallee team that did the original work on Libon. Normally it's the enterprise unit that's most empowered to do its own thing at a telco or some of the web/content guys, but hopefully with DT its various national OpCos & T-Labs can exert more leverage as well.

Monday, August 10, 2015

Trip report: a tale of mobile/WiFi in two developing countries, Haiti & Cuba

I've been away for the past 3 weeks in two very different, but very close countries: Haiti & Cuba, separated by less than 100km. While it was a vacation and I was mostly "off-grid", there were still a few interesting things I noticed about local use of communications and the Internet in each place.

Haiti is one of the poorest countries in the Western hemisphere, and has some of the worse conditions of poverty and restricted infrastructure I've seen anywhere. Outside the Petionville district of the capital Port-au-Prince, there are still many signs of 2010's devastating earthquake, and often severe poverty. It's a beautiful country, with some fascinating places - but it's also one of the hardest places to travel that I've visited.

However, cellphones are pretty ubiquitous, with growing use of low-end smartphones, but still a lot of basic voice/SMS devices around. There are two mobile networks - Digicel, which is prominent throughout the Carribean and various other island nations, plus fixed/mobile Natcom which is majority owned by Vietnamese telco Viettel, but apparently holds a small market share. Most users (except roamers and wealthly UN/NGO/government types) use prepaid SIM cards, with top-ups available from many locations, often street-side or even sold through bus windows during stops. Adverts for mobile networks are everywhere, often permanently painted onto walls of houses or shops.



3G coverage is present across a fair amount of the larger Haitian cities and by major roads - I visited Jacmel and Cap Haitien as well as the capital. (By coincidence, one of the press releases in my post-holiday inbox is from Astellia, talking about a contract with Digicel for 3G network optimisation in the country).

Data charges are reasonable to external eyes (eg 20c/day for 90MB, or bizarrely at an anti-discount $8/mo for 2GB - see here) but still pretty expensive for many of the inhabitants (my guide used a BlackBerry because of its compression abilities, primarily for email & Whatsapp), but seem to be of growing importance to many. I noticed various Facebook-centric per-app or zero-rating plans being advertised.

The few good hotels in the country typically had decent fixed broadband and WiFi, but I saw very little of this in other locations, fairly unsurprisingly. Interestingly, Haiti is included in Vodafone's £5/day WorldTraveller programme, so I could use my iPhone in most places at relatively reasonable cost, rather than getting a local SIM for my spare phone - although had I known the 20c/day rate and been staying more than a few days, I would still have taken that approach.

My overall sense was that Haitian use of mobile and the Internet is broadly on a par with other developing nations at a comparable level of GDP, and apart from patchy competition it seems to have similar business and deployment models. I noticed various schools offering IT and Internet lessons, although some of the poverty I saw suggests that adoption across the whole population will be slow - there are more important problems to fix first.

Cuba, by contrast, could not have been more different.

Firstly, I switched off data roaming as it's not covered by a decent plan, but would have cost me £3/US$5 per MB. Roaming for phone calls & SMS seemed to work OK in some places but not others - I had three days of "No Service" in the middle of the country. One SMS I sent to a local Cuban took 4 days to arrive. Apparently there is 3G data available in some places, but it's aimed at tourists rather than local inhabitants. Relatively few Cubans seem to have phones anyway - although that is changing (see below). It's the only country I've visited recently that has payphones everywhere on the street - and people using them.




It seemed to be possible for tourists to get local SIM cards (eg to call hotels or the small number of private restaurants), but I didn't bother as I suspected it would mean navigating assorted bureaucracy in Spanish. I did, however, get a calling card for the payphones - the first time I've used one in about a decade. There is one state monopoly provider of fixed and mobile communications, ETECSA, and I didn't see the proliferation of mobile-related advertising you get in almost any other nation.There's quite a few places that are agents for top-up cards or payphone cards, but they're not the "phone shops" you'd get elsewhere.

I'd already told all my friends and clients to expect me to be offline, as I knew Internet access was near-impossible to find except on PCs in hotel lobbies, or telecom operator offices where I'd likely have to stand in a queue in the heat outside for ages. I'd heard that prices were $4.50-$10 per hour - well outside the reach of most Cubans, for many of whom (under the dual-currency system there) that would be a sizeable % of their monthly wage packet. Looking at some guides & reports online, I found that a few locations had WiFi, but they were mostly in Havana rather than country-wide (I visited about 7 different places). There are very few URLs displayed anywhere.

But that has changed, and very recently.

On July 1, ETECSA cut access rates by 60%, to $2/hour (bought via scratch-cards from their booths and offices, or some hotel lobby desks). They also fired up WiFi access points in various public squares and parks - often the social hubs where hundreds gather in the cooler evenings anyway.

As a result, when I visited 3 weeks later, I saw small clusters of both tourists and Cubans sitting in the shadier bits of the squares, clutching phones and tablets, wherever was closest to the telecom office or WiFi AP. I saw mostly cheaper Android phones and no-brand tablets, but quite a lot of Samsungs and a few Apples. Amusingly, some devices sported US carrier branding, suggesting recycling/unlocking of old phones. 

The WiFi (still expensive by local standards, but about the price of 2 cans of beer), was actually pretty speedy. None of the services I normally used seemed to be blocked, either.

But the most amazing thing was the realtime behavioural change I could witness. There were huge queues outside all the ETECSA offices. Out of hours (or alongside the queue) some enterprising folk re-sold the scratch-cards at a markup for the impatient. Some hotels ran out of stock of the cards, which could be used at the (ETECSA-run) indoor WiFi as well as the public venues. Online access had very quickly become the teenagers' entertainment of choice.



And then I started seeing cross-generational groups standing on the street in the evenings, using video-calls (Skype or Viber?) to connect kids and grandparents to relatives elsewhere in the world (presumably many in the US, to which numerous Cubans had emigrated during the last decades).

It will be very interesting to see what other societal shifts occur in coming months in Cuba. The Internet genie is very much "out of the bottle" - unlike a lot of countries, the low pre-existing use of the web means that the first use for many is on a smartphone or tablet. And, interestingly, via public outdoor WiFi, rather than cellular or a fixed PC. 

Ironically, the only city where I couldn't easily find any public WiFi was Varadero, the package-tourist capital of the country. And in Havana, I couldn't find a square with "vanilla" ETECSA WiFi access, so I instead ended up using the networks at hotels or Hemingway's favourite daiquiri bar, Floridita.