Pages

Pages

Thursday, July 26, 2012

UK Net Neutrality - moving towards the right idea

I've long taken an interest in the machinations & politics around Net Neutrality, especially here in the UK. My view has long been that fixed and mobile operators should be allowed to sell various "managed service" forms of data connection and broadband service, as long as it is made absolutely clear to end-users that this is not "Internet Access", and that it doesn't unduly interfere with Internet Access also provided on the same infrastructure.

My view has been that "Internet Access" means real Internet Access - ie as neutral and non-discriminatory as possible, except in emergency circumstances relating to necessary traffic management to protect overall network integrity.

(I'm not too purist about Net Neutrality, because most of the more slanted and anti-OTT vendor notions about application-based policy either don't work technically, or wouldn't be accepted by users in competitive markets)

I've talked about an "Appelation Controlee" or maybe a protected brand for the term Internet. It's like Champagne - it only comes from one place. Other drinks can't use the name. But it doesn't stop your local bar selling Cava or other sparkling wine as well.

If you're an operator selling Internet Access (TM), it has to be 100% Certified Prime Internet Access. Not "processed Internet-like substitute", or "I can't believe it's not Internet". You can call that something else, but it needs to be clear to customers that's it's fake little-i internet and not the real thing.

All traffic management policies should be made transparently available, and regulators should encourage the use of neutrality-checking tools by users, to make sure that they're not being hoodwinked. It should be a criminal offence equivalent to fraud, to deliberately or covertly manipulate open-Internet Access traffic.

Users should also be protected against non-Internet managed services causing congestion on shared links, to the degree that full-Internet becomes useless. Publishing actual performance targets and attained historic numbers, rather than headline speeds, would be a good way,

It should then be up to a government & population to decide whether offering some form of Internet Access should be a universal service obligation, and based on what criteria (eg average attainable speeds & latencies).

After that, operators should be free to experiment with different models - partnering with software or Internet companies on new products, developing new charging models and so on. Ideally, this would be separated from proper Internet traffic - perhaps even using different frequency bands on cables or airwaves, although that will not be universally feasible.

Two years ago, I wrote up a suggested "Draft Code of Conduct" that gave some guidelines that I felt were a reasonable compromise between a hardcore NetNeut position, and a completely "liberal" telco-centric "we can do what we want with our pipes" viewpoint.

I then polished this up into a submission for UK Regulator Ofcom's 2011 consultation on the topic, which is downloadable from here. (The submission has slightly less opinionated views on what I thought "should" happen than my Draft Code)

Then last March, a group of ISPs issued their own first attempt at a Code of Conduct. My discussion of it is here. I felt that it was an OK start, but tried to treat traffic types (video, text etc) as applications, which is wrong - video is 1000 different applications, often embedded into other applications. Trying to separate out an arbitrary traffic type for different treatment is like charging more for a magazine depending on the amount of green ink.

The next step in the saga was Ofcom's publication of its stance on neutrality in November last year. This very clearly laid out the distinction between Internet Access and managed services:


"In particular, if ISPs offer a service to consumers which they describe as ‘internet access’, we believe this creates an expectation that this service will be unrestricted, enabling the consumer to access any service lawfully available on the internet. As a result, if a service does not provide full access to the internet, we would not expect it to be marketed as internet access."

This fits very closely with the view I've been propounding. It doesn't stop operators offering different sorts of connectivity - for example, prioritised IPTV or VoIP. It does however mean that any partner-type models involving end-to-end prioritisation (ie on the access network, not just middle-mile shortcuts like CDNs) would not fit within normal Internet access models, but would have to be marketed separately. This is another nail in the coffin of unworkable concepts like 1-800 Apps for the UK.

This was then followed up with the European Regulators' group BEREC announcing its preliminary findings on traffic management earlier this year (Spoiler: lots of places block VoIP) and it is now doing a full consultation up until October. It too has noted the difference between real-Internet and fake-Internet services, and is considering options for allowing prioritisation, blocking etc.

The Dutch, meantime, have introduced "hard" Net-Neut legislation, following KPN's rather careless remarks about blocking or charging extra for WhatsApp or Skype. Other operators have suggested some form of Internet tax to protect their legacy revenue streams from others who've actually innovated over the last 100 years since the invention of the phone call. (This is the usual weak argument of "we're dumb, so can you please tax the clever people for us").

Now, a new version of the UK ISPs' code has been issued - the full document is here . This is the pivotal line, in my view:

"Signatories also recognise the importance of best efforts internet access being a viable choice for consumers alongside any managed services that might be developed and offered."

As well as:

"in instances where a product does not support full internet access, i.e. where certain classes of content, applications and/or services are blocked, the term “internet access” will not be used to describe or market such products. ISPs also commit to effectively communicating any restrictions on such products."

(There's a bit of semantic wiggle-room here, with "blocked" not necessarily including "degraded", but I suspect Ofcom will come down like a ton of bricks on anyone following the letter and not the spirit of the thing, not to mention considerable PR and competitor opprobium)

The stated ability to do traffic management to avoid serious congestion, block child-unsuitable material, avoid DDoS attacks etc also seems reasonable.

Supposedly, about 10 ISPs have signed up, but reportedly Vodafone hasn't because it thinks it's too restrictive about the term "Internet Access", and Virgin Media thinks its too laxly-worded and wants loopholes tightened up.

In my view, Vodafone: you need to grit your teeth and sign. Internet is Internet is Internet. If you want to sell something else, I'm sure your marketing  department is smart enough to come up with another brand. I can even give you some advisory input if you get in touch. (How about "Internet-flavoured access"?)

Otherwise overall, I think the UK ISP community is now world-leading at developing a suitable compromise between Net Neutrality, leaving scope for alternative telecom products and new business models.

I'm pleased that I may have helped, in a small way, to clarify some of the arguments and debates with my various posts and official consultation submissions.

Monday, July 09, 2012

What's your "ecosystem mix"?

One of the most interesting big-picture trends in telecoms at the moment is what my colleagues at Telco 2.0 / STL Partners refer to as "The Great Game" (A historical reference to the Russian and British Empires' strategic struggles over control of Central Asia in the 19th Century).

Which of the various "ecosystems" will win the hearts, minds and wallets of consumers? Telcos, Apple, Google, Microsoft, Facebook, Amazon and so forth are in the frame here. The empires concerned span access, devices, voice services, messaging, social networks, advertising, content, apps, payment mechanisms, customer data, authentication and more.

The more components in each ecosystem, the greater the potential for lock-in. Apple is hoping that iPhone owners will buy iPads and Macs - and use iCloud and other services. Android users are steered towards Gmail, Google Apps and perhaps even Chrome on the desktop. Mobile operators are hoping to maintain their role in voice telephony, and add social functions, content and payments. Facebook's reach is extending across the web and betond - most recently, even to owning sub-oceanic fibre.

Some people have already made their initial decisions - there are Apple fan-boys, Google enthusiasts, quad-play telecom subscribers and so forth. Others are adamant about which companies they want to avoid - for example, there are many Facebook refuseniks, as well as people whose only telco interface point is via a SIM and anonymous prepay top-ups.

Myself, I fall into the camp of "a plague of all your houses". I deliberately structure my use of technology and services to avoid becoming beholden to any one of these organisations. I don't want to be bombarded with adverts, I don't want to be locked-in to any platforms, and I certainly don't want complete profiles of my life being deduced by any analytics systems. Paranoia or just sensible precaution - who knows? I also don't want to become a digital recluse, so my answer is to use a "bit of everything".

In particular, I try to keep work & personal stuff separated, except through touchpoints such as Google Search or my home broadband. (Hmm.... maybe I should delete personal friends from LinkedIn to keep it "pure" and business-focused).

My basic premise is that I don't want any one provider to see or control more than 20-30% of my communications & Internet existence. This is one reason why I wouldn't have an Android phone, and if I'm using an iPhone why I wouldn't also want a Mac - and why Apple doesn't have my credit card details in  iTunes. Google for me is about search, maps, this blog and an Adwords account. I sign out when I can, and wouldn't touch G+ if you paid me (and the only reason I use G Docs is for communicating with clients who do pay me).

It's also why Disruptive Analysis doesn't "do" Facebook, why Twitter is strictly work-only, why I wouldn't use a mobile phone account to buy stuff, or ever sign up triple/quadplay services. It explains why I had to create a new random Microsoft Live account to test a Nokia Lumia. It's also why my main personal email is still a (paid!) Yahoo premium one. I get music from Spotify, VoIP from Skype, books from Amazon and happily use Paypal and credit cards for most online transactions.

Frankly, the best thing about the Internet is choice and divergence not convergence into monolithic ecosystems.

What about you? Have you sold your electronic souls to any one of the communications devils? Or are you determined to withhold "loyalty" until it's is actually earned? And am I lulling myself into a false sense of security?

Monday, July 02, 2012

New Report: 10 Reasons Why the "toll-free" 1-800 Apps Concept Won't Work


I've published a 35-page report on one of the mobile industry's lastest overhyped ideas: the concept of a "1-800" model for apps, where the user doesn't pay for data against their quota or cap, but the app or content publisher picks up the tab instead.

The idea has been floated by various operators - notably AT&T and Verizon Wireless - and supported by assorted network and policy-management/charging vendors.

Although superficially compelling, there are numerous practical flaws with the idea, ranging from difficulties with pricing of such "sender-pays" data charges, through to unexpected side-effects as users and developers try to "game" the system. Add in uncertain IT systems, the rapidly changing nature of apps themselves and assorted other "gotchas" and the idea starts looking unworkable.

The report elucidates 10 separate reasons why the general model for toll-free apps can't work, as well as identifying a handful of niche use-cases where it might be OK.

One of the main problems is defining exactly what an app actually is - I've mentioned the famous YouTube-video-inside-Facebook example for a couple of years, but it's now getting much more complicated than that. HTML5 means that "apps" can be created on-the-fly, individualised for different people. How does a 1-800 model work when your Facebook app and mine are different?

Then there's things like WiFi. If I'm an app developer, why should I pay for someone's mobile data traffic when I can just suggest they connect via WiFi instead? Maybe for the small amount of "absolutely need-it-now" data there's more of a case, but that makes it much more difficult to separate out the time-critical vs. delayable parts of the app. Unless the price is so cheap that the developer doesn't care - in which case the operator probably won't be making much money either.

Then there's the IT and back-office side of all of this. How does an app developer sign up, track their data costs, manage fraud and abuse - or complain if it goes wrong and they're mischarged by the operator? And if you try and implement a version of the tollfree data model where the traffic isn't just charged differently, but prioritised or given extra QoS it gets 10x harder still.

My view is that it's a nice idea for press releases and conference soundbites, but when you look at the detail it just won't work, especially in the general case where any app-developer or content provider can use the platform. The 1-800 phone industry - where anybody from a florist's shop to an international airline can offer free calls - simply cannot be replicated for mobile data and apps. There are far too many gotchas, many of which I'm not mentioning in this blog post.

Let's just work out a way to gracefully back away from the idea, look at a couple of niche and highly-customised special cases, and move on. "Sender-pays" for mobile data simply will not work for massmarket apps and content.

The report's contents (total length 35 pages):
Introduction       
Several different models of “toll-free”   
1. No clear definition of “an app”          
2. Business Model Fit   
3. Apps’ data usage is device-dependent
4. Apps have multiple functions      
5. HTML5 changes the game again  
6. Poor fit with WiFi and femtocell offload  
7. OSS/BSS challenges & “dashboards
8. How would operators price toll-free? 
9. Network dependencies & standards   
10. Arbitrage & other unintended consequences
Can two-sided business models ever work
Exceptions where Toll-free might work
Conclusion     
About Disruptive Analysis

I'll put up a more full blog post and description in coming days, but for "early adopters" willing to pay upfront, I'm offering a $100 discount for a short introductory period.

The report is available as a PDF only, delivered via email within 24 hours of payment via the Paypal / Credit Card link below. The discounted price is US$595 for a 1-3 user licence, and $895 for a corporate licence. (If you're in the UK or EU, you'll need to select the option with 20% VAT added).

If you prefer to order and pay via PO/bank transfer, please email via information AT disruptive-analysis DOT com . Also please note that I have been experiencing occasional problems with the Paypal BuyNow links below - please let me know if you have trouble.
  
If paying via a personal email account, please also send me your work email address or a contact phone number.


Disruptive Analysis 1-800 Toll-Free Apps report:
1-3 User Licence


Toll-free apps report (PDF) 1-3 users

Corporate Licence


Toll-free apps report (PDF) Corporate