Signaling Blog

Forget iPhone, the Gphone is here

One of my favourite Martin Geddes quotes is this one:

Here’s the deal. In the old world, the economic activity started when the phone call began. In the new world, that’s when the economic acivity ends. The money is all in presence, social networking, filtering, privacy management, and so on. It’s a complete inversion of the economics of telephony. Therefore expect many of the vendors to be disemboweled in the process.

Although no man can predict the future with complete accuracy, I would whole heartedly agree with the sentiments being expressed. So I was quite happy when trawling back thru two months of unread blogs to discover not only a good Telco 2.0 article but another great quote in a similar vein:

The basis of competition in this third wave is about prompting and connecting people who need to interact — social and contextual awareness — whilst preventing unwanted interruption and connection — i.e. security and privacy. By the time the phone call starts, the money-making is over. If the call was wanted, it’s the broker who gets the payment; if it wasn’t, it’s the filter.

Today I came across the Triple-I conference (I-Know, I-Media, I-Semantics) which takes place in Graz (Southern Austria) 5-7th September. All being well I plan to attend. I am glad to see such a multi-disciplinary approach. I certainly see great overlap and synergy between knowledge management, new media and semantic web technologies. Personally I’d opt for a Quintuple-I conference so that a ‘I-Communications’ and ‘I-Positioning’ was included!

Skype has been down for well over 24 hours now. In the meantime I switched over to using GizmoProject - shame my contact list on Gizmo is so bare. For those who do not know, Gizmo is a Skype clone; it is powered by the same great Global IP Sound (GIPS) codec as well as also using a self-organising peer-to-peer (P2P) network to route the call signalling. It tries to differeniate itself by claiming to be based on completely open standards (e.g. it purports to use standard SIP, although I am unsure how the P2P element is then handled as it is not using any P2P SIP standard) unlike Skype which went for a blackbox approach. I’ve often had software crashes and other problems with Gizmo (C++ runtime errors etc.) plus I dislike the interface in comparison with that of Skype - let alone the fact that my buddy-list is well populated on Skype. Hence when Skype eventually works again, I will be switching back.

For those battling with the telecom vs Internet reliability dichotomy I’d like to throw in a point to reflect on. Signalling system #7 (SS7/C7) has had its degree of outages. See for example Aspects of Integrity in the NII , John C. McDonald ,MBX Inc. Quote:

“Several well-publicized SS7 outages occurred in 1990 and 1991 due to software bugs ^{[6, 7]}. The first had a nationwide impact and involved the loss of 65,000,000 calls. Others involved entire cities and affected 10,000,000 customers. In response to a massive outage in September 1991, the mayor of New York established a Task Force on Telecommunications Network Reliability. The task force noted that “the potential for telecommunications disasters is real, and losses in service can be devastating to the end user” ⁸.”

6 Fitzgerald, K. 1990. “Vulnerability Exposed in AT&T’s 9-Hour Glitch,” The Institute, March.

7 Andrews, E. 1991. “String of Phone Failures Reveals Computer Systems’ Vulnerability,” New York Times, July 3.

8 City of New York. 1992. “Mayor’s Task Force on Telecommunications Network Reliability,” January

This was sent in and I have posted it here as it gives a rather coloured and yet refreshing angle on the issue of quality of service (QoS):

Back-drop

“So what” …is Quality of Service (QoS)? The constant use of the term reminds us of the Chicken Little cries that the sky is falling …but this time it is true in the case of the ITU circuit based business models! So let’s start with the circuit based business model. Recall that a circuit is a particular bit rate; say 64kbps – no more and no less. For decades Internet Service Providers have been buying circuits and putting packet switching gear at the circuit termination points and then sending connectionless IP packets over it. Voila, the Internet!

Today the entire planet is getting on the IP packet “OPEN” Internet bandwagon and its enabled cornucopia of services. In fact by being OPEN the Internet has enabled the greatest creation of shareholder value in shortest time in history! Of course the operators have noticed, but don’t want a pieve of the action they want all the action. Their problem is “how to co-opt the Internet?”. Since there are NO circuits within the Internet, how does the telephony cartel recreate circuits to reinvigorate their business model and control all the new cool services emerging from the Internet? The answer is a fictional problem solving but user abusive rate limiting mechanism called QoS that allows the Telco’s to provision traffic flows that enforce a particular packet bit rate; say 64kbps – no more and no less (see “MaxBitRate” definition in 3GPP TS23.207). It’s obvious: the telephony cartel wants the world to evolve from the packets over circuits infrastructures of today to circuits over packets infrastructures tomorrow, enabled by QoS. But where is end-user value in the operator plans?
The IP packet technology driven services situation today is painfully obvious to the Telephony cartel. The Internet has hugely succeeded and IS our future in spite of the fact that no one in the telephony world, for decades, believed in its connectionless packet based protocols and technology. An open Internet giving each end user freedom to access the services he wants and the huge constantly expanding value delivered by the Internet have backed the telephony crowd into a corner and now they are desperately trying to CLOSE the Internet and turn it into something they can control! QoS is a lynch pin of that strategy.

Recent telephony cartel activities focused on rolling back time to the circuit world of yesterday (prior the popularity of Internet) are rooted in the 3G and IMS standards that are part of the general ITU, and now ETSI, reaction to inspect, control, inhibit and charge for all data traffic, and therefore “own” the Internet. The clear telephony cartel direction is to create a walled garden environment for application development and combine that with levying toll charges, via QoS, for end users to get outside the walled garden. This one two punch is expected to fend off future as well as existing Internet based threats such as Yahoo, Google, Skype and others.

More specifically, QoS from the ITU and ETSI cartels is a reaction to the immediate VoIP (Skype, Vonage, etc.) threat to existing telephony business models. Because voice is today’s killer application it has become the first battleground for operator monopoly business model protection. Expected data revenues are not materializing to off-set declining voice revenues. Note that 3G might be “almost” here, but it has failed miserably in achieving its data business objectives. 3G-only operators continue to struggle to find a workable business model and continue to compete on the declining price of voice. The sooner the telecommunications operators realize the need to start driving towards their final bit pipe future, the sooner they can halt their wasteful spending during their inevitable decline and ensure their survival (just look at the pathetic history of AT&T).

Pipe Dreams?

Will QoS, without assistance from today’s regulators and political lobbying, help the Telco industry survive the Internet assault on its business model? Technically No! – not directly, because unfortunately operators and their associated vendors are learning that voice is highly compressible, and so low in bandwidth (voice traffic is negligible on WANs, and as LANs as well) that customers can use it over today’s Internet without any Telco operator involvement.

The real problem is no VoIP revenue flows to the operator. Instead, I suspect paying for QoS guarantees will be forced on the customer to bypass secret business model protection plans sometimes called “p2p mitigation”; In other words, artificially created IP traffic disruptions. The general idea of these business model protection plans are to add latency and obstructions to network infrastructures to cause quality problems for Skype, AOL, Apple, Google, IBM, MSN and Yahoo (they all have voice as a free add-on) and/or other VoIP activities and in general prevent competition via the Internet protocols unless you pay more for bypassing these artificially constructed latency or congestion constraints. Sometimes this situation is called traffic “modulation” which is really the creative introduction of jitter into traffic flows. This behavior is similar to gang activity that will leave your business alone when you pay protection money.
As with QoS there are always “corner cases” of abuse that p2p mitigation advocates will try to generalize across the entire Internet population. The p2p mitigation advocates are implying we are all bad (illegally downloading music, videos and other dubious content). Therefore we need to be “mitigated” and also pay for the privilege of being “mitigated” because we are likely to consume the advertised bandwidth we actually paid for with our illegal activities. But you should note that p2p activity includes Skype and other valuable edge to edge services which is something p2p mitigation advocates conveniently fail to mention. The claims of illegal activity are just another thinly veiled excuse to exercise more control over Internet without any end user benefit.
“p2p mitigation” is what I call QoS+EL = Quality of Suppression with Enhanced Latency. Available bandwidth, latency and packet loss can have significant impact on the perceived quality of any service. With today’s technology it is a small step for an operator to badly impair traffic unless you give in to their extortion demands for QoS protection. The idea is to provide a premium service and the appropriate premium will be set by how much pain end users would suffer without the QoS premium service.

We have already seen service providers blocking Vontage and their anti-competitive extortion motives are clear. The most plausible future scenario for end user abuse is an IMS or TISPAN operator blocking, interfering with traffic and/or providing lower quality connections for Skype traffic or to Google or any other “outside the walled garden” service on the Internet. “p2p mitigation” by any operator without the user being informed and gaining his consent seems to be an obvious abuse with the malign intent to extort money from the end users (at both ends of the connection), via QoS, as well as prevent competition. In summary, QoS strategies fall into three categories (1) a DOS attack on customers who don’t pay extra, (2) to extort content companies, (3) to suppress competition.

Moore is Always Better

Looking back over nearly 50 years of networking technology progress, the networking industry (following Moore’s Law) has always improved bit rates and consistently reduced network latency seen by applications and the end user. This is simple to do and is a good thing as “speeds and feeds” continuously improve. In fact, light wave and radio technologies are adding capacity even faster than Moore’s Law predicts today!

In the past, Operators just blindly following their business model and provided connectivity. They didn’t pay any attention to what it was being used for, and never really looked for better ways to serve the public. So over time, improving capacity and reduced latency allows cool new things to come out of the Internet eco-system and this almost always bypasses the operator. Why? Because operators are simply NOT networking people they are telephony application people that just sell switched circuits. They have blinders on and can’t see technological or networking change coming, when they do, they react too slow, or react with disbelief as they continue with insulated group think activities at the ITU-T and ETSI lemming races.

Unbelievably, the telecommunications industry now wants us to believe that they need QoS because Moore’s law isn’t providing capacity fast enough, which we all know is nonsense. So out of one side of the Telephony face they imply Moore’s Law is broken when it comes to capacity, while the other side of their face seems to be quite happy proposing they use all the on-going new technology break-throughs (thanks to Moore’s law) to do real time deep packet inspections that enable controlling content access, micro-bill, dynamically reduce bit rates or increase latency and/or construct impediments to end user traffic flows. The rational for QoS is clearly backed up by hypocrisy with a topping of ultra-thin marketing arguments. But today’s reality of a vibrant Internet eco-system without QoS clearly shows the lack of any need for QoS. If QOS really did something useful, like providing end user value, then the economics would have forced it into widespread use by now. Let’s face it, the only ROI for QoS is extortion! In fact this pathetic attempt at controlling the Internet while extorting the end user or content providers is really hard to do, complex, expensive, and has ramifications on innovation and the future health of the Internet.

Drinking Kool-Aid at Camp

Today there are two QoS camps. Which one do you think is looking out for the end user?

• People SELLING gear and SELLING services based on that gear to the operators for imaginary, but revenue generating, quality problems. One characteristic of these QoS religious zealots is to find a corner case, like a 256 kbs link to a distant land, and tout them as the main situation. Without artificially constraining bandwidth, today’s QoS supporting arguments seem more like religion and are simply not convincing. The zealots are simply trying to sell QoS gear and traffic shaping devices. It’s just sly marketing.

• People who like the simplicity, honesty and freedom of the sticking more closely to the architectural principles of the existing Internet. The simple fact is the Internet has become extraordinarily valuable to everyone, except the carriers. The Internet generates enormous wealth and associated tax revenue. The Internet is extremely useful for information and a key part of the global economy. The fact that it is hugely important to so many people without QOS says volumes about those pushing the QoS Kool-Aid.

What Does the “I” in “IETF” Mean?

Can we look to the “ITU Engineering Task Force” for help? Many believe the IETF needs to do a deep dive on the question whether QoS is desirable or not and if so, what form it should take. Remember that QoS does not create bandwidth and is not a substitute for bandwidth. So investments in bandwidth will still be necessary. But, today too much of the IETF is moving forward with the assumption that complex expensive QoS is needed in lieu of simple increased capacity. This flies in the face of Internet history which has taught us that as capacity increases new ideas and innovation takes root and the Internet increases in value. An ITU-T (carriers and legacy vendors) revenue agenda is behind the motives of those trying to move us away from this tried and true formula of Internet success.

In fact better solutions than QoS exist today! For example, DiffServ applied by the end user can be effective and lightweight (much less complex) to implement. DiffServ allows all the different
traffic types to share the same link and assure that each traffic type enjoys the bandwidth and latency required for it to provide a good end user experience.

If DiffServ works then why use QoS mechanisms that are very complex, extremely awkward, heavy-handed and prone to abuse? QoS discussions deserve a much broader treatment in the IETF and the IAB. But I am especially concerned about the increase in IETF attendees that:
- are not at the IETF as individuals but instead “hum” the company revenue tune.
- don’t have product responsibility. Instead they just write theoretical RFCs designed to pad their own resumes or get brownie points in their companies. - run back and forth as errand boys between the IETF and the ITU-T, ETSI, OMA and any other Telco dominated standards body or consortium (most notably 3GPP and 3GPP2), - don’t understand the Internet (especially the cellular people) and may be openly hostile to important Internet guiding principles like providing end user value and non-interference to the end user ability to consume the services they want.

This brings up a significant concern. Will the current mix of the IETF attendees result in the proper outcome of any QoS debate? Is the IETF up to the challenge to develop a mechanism or policy to fix this, and at the same time successfully deal with Internet neutrality and phantom QoS benefits incessantly hyped by marketers?

Share the QoS Love

Many of us heard about the IETF “love fest” at the ITU-T a while back. In fact, the love can still be felt as 3GPP meets at IETF meetings with IETF leadership with non-advertised regularity. What ever happened to the idea of IETF transparency and more importantly being a gathering of individuals and not pandering to organizations? …you know “…no queens…” and all that rhetoric? Today many of the IETF working groups are just doing work for 3GPP and other ITU-T bodies. If you require more convincing of the cozy ITU-T and IETF relationship please check the IETF web site (http://www.ietf.org/meetings/IETF-xx.html, where xx is the latest IETF meeting number, try 64 or 65 for example) for a list of host and contributor companies for the latest and previous meetings. You can argue against this perspective but many have been there …and that perception is reality!

Sales and Shills

I am afraid the IETF has become a shill for the ITU-T and 3GPP/3GPP2 by adopting the ITU mindset and a telephony cartel revenue agenda simply because they are being over run by Telco people? For a few IETF data points (and there are lots more), consider the ongoing lemming race for complexity between DIAMETER and RADIUS proponents to see who can best monetize services and micro bill the end user. Also consider proposals at the “Layer 2 control mechanisms BOF” to standardize additional complex end user rate restriction mechanisms. Notice that new activity for peer to peer SIP is attempted but constantly delayed in spite of clear support in IETF BOF sessions. IETF: WHERE is this needed SIP P2P WG?

But isn’t SIP already peer to peer? It seems telephony restrictions on SIP have resulted in Skype envy and renewed the desire to get out from under ongoing telephony suppression of SIP flexibility. But this doesn’t stop the SIP and SIPPING WGs from their not so well hidden agenda of endlessly setting up billing possibilities for Telco’s. For example, sipping is proposing SIP changes that supposedly allow end users to accept or reject content transfers but have the real motive to facilitate operator billing and then use the excuse of “managing” SIP entities behind NATs and Firewalls to accomplish a form of electronic home invasion (other WGs such as hip, have similar agenda’s). Next, look carefully at all the past and present IETF and Internet leaders now working in or beholden to the ITU or Telco cartel eco-system for revenue. Listen carefully to whether these Internet pundits support the end user, avoid the QoS debate so they are not exposed as hypocrites, evade QoS questions possibly due to pangs of conscience, or simply find excuses to follow the telephony cartel revenue agenda so they can enjoy the rewards of hobnobbing with bureaucrats at the ITU as they abandon founding Internet principles.

The IETF is at a cross roads and must decide either in supporting the end user or driving a revenue agenda. This boils down to two crucial choices. The first choice is between (a) simple standards developed for end user value which is root of the IETF success or (b) complex standards developed for Telco cartel restrictions and revenue agendas. Note the revenue agenda for creating standards have been an ITU specialty with well a known result: complex expensive systems that fail expensively in the market. The other key choice is between accepting standards written (c) by product developers or from a research project …you know “running code” or (d) by professional standards meetings attendees which is another ITU specialty.

I think that many of us are worried that QoS just represents another example of the IETF leaning towards the telephony cartel strategy of developing standards for a revenue agenda, market control and innovation threat suppression. Just look for innovation at the IETF today. Do you see any? Has it stopped because the Telco money at the IETF wants a closed system to suppress innovation threats and provide for business model protection while aided by the short term Wall Street focus of the technology sector? There is not much time left to correct the situation before the IETF goes completely over to the ITU dark side. But I have to wonder if the IETF can return to its core belief that successful standards are made for the end user. The ITU has already learned that end users always vote with their wallets and losing standards are made with the agenda to make money without caring for the end users. Is it necessary for the IETF to also learn the same painful ITU lessons?

Those supporting QoS make a lot of noise, but real support for QoS is far from convincing. Notice that not everyone has drunk the QoS Kool-Aid. All the public VoIP providers such as Skype, Vonage, and Google etc. are doing just fine without QoS. Please note that public VoIP providers can’t use QoS, even if they wanted to, since they don’t control the network. In fact, no single VoIP provider (including those that do own the network) can control QoS end to end anywhere across the world. So why try? …except to exploit its rate limiting date flow suppression properties. For an example of the lack of need for QoS, consider earlier this year the U.K. carrier British Telecom announced that it intends to target its BT Fusion converged service at larger businesses and offer VoIP calling when WiFi-enabled handsets become available. QoS wasn’t even mentioned and it is not a big deal for BT. Statements were made that transmitting VoIP over WiFi hotspots isn’t that big of a deal. They just transmit VoIP traffic like any other data. What NO QoS!?!! …Gasp!

Quality of What?

What is the real QoS (+EL) motivation? I think the answer can be found in the numerous interpretations of the letters “QoS”. Consider the following meanings of the true definition of QoS by analyzing all the alternatives the acronym really stands for.

• Quality of Suppression. Let’s face it. QoS is just a rate limiter. The term QoS is a great marketing ploy to “infer” that some kind of “quality” is provided to the end user. This is a great way to deceive and simply build rate limiting mechanisms. Since rate limiting is the case, then Quality of Suppression should be the proper label. But you know operator and vendor marketing departments will never go for that! But this is simply truth in advertising. Consumers pay for Internet access and bandwidth and also expect “neutrality” which allows usage of the bandwidth they paid for along with unimpaired access to any resources on the Internet.

• Quality of Stupidity is the propensity of the IETF, government regulators and uninformed end users to swallow this QoS FUD and not puke. Telephony operators and vendors have come to rely on the “masses are clueless asses” strategy as a key means to protect their empire and financially stick it to the end user. This ITU strategy is being applied without prejudice to the IETF, world governments and all end users. How long is the end user going to accept the short end of the stick and both the IETF and governments continue allowing them to be restricted and financially abused?

• Quality of Secrecy is all about getting billed for advertised capacity and really not having it. The operators are on a path to deliver as little value as possible to the end user and work diligently in the background to construct impairments and other obstacles for service access outside their walled gardens (for paid QoS to eventually alleviate).

• Quality of Surprise is the look on the end user face when things are slower than before or his voice or video over IP sessions stop working. Of course this can be turned around by paying for QoS and then asking your carrier to prove to you that you are getting the QoS they promised. Then look at who is surprised!

• Quality of Shocked occurs when you find out the Telco’s are still trying to protect a 100+ year old monopolistic circuit based business model and the end user is paying more for less.

• Quality of Suck is the sound of your money is being pulled to the Telco’s without providing any value. This behavior can also referred to the Quality of Separation, as money is separated from end user wallets without any value being returned.

• Quality of Senselessness is the result of the ITU suppressing innovation and the occurrences of exciting new services as well as reducing overall wealth and world wide tax revenues from Internet activities. Hasn’t the ITU figured out that an OPEN INTERNET generates fantastic amounts of wealth and tax revenues?

• Quality of Silliness is the activity focused on protecting a dying circuit business model and then perpetrating monopoly abuses, such as QoS, against the Internet end users.

• Quality of Scared is the operator and ITU reaction to the inevitable conclusion of becoming commodity bit pipes. On an up note, their desperation in the marketplace, standards bodies, regulatory venues, press, legislatures and the courts should be “fun” to watch

• Quality of Stunned is the inevitable vendor reaction to the drop in sales and stock prices as they continue to perpetuate the fantasy that QoS does anything useful for the end user and to the ITU lemming crowd that continues to build complex and expensive gear no one will buy.

• Quality of Sanity is for all those who challenge the QoS myth.

Special Packet Handling

QoS is an extremely complex subject and I can’t do it justice in this short article so please refer to any number of appropriate texts for more detailed information. So rather than getting into details, I’ll focus on the main problems of rate limiting and traffic modulation by QoS mechanisms. Rate limiting and traffic modulation is what is the most attractive to the Telco cartel since these mechanisms can be used to disrupt outside the walled garden services and also dial back bandwidth. Reduced bandwidth, in particular, creates artificial constraints. QoS can then be used to shake down any end users who wish adequate Skype or other p2p services.

If QoS marketing shills were truly worried about the end user experience, and not just focused on inward looking business model protection, then the recommendation they should make is for better queue positioning mechanisms for network traffic classes and not rate limiting by individual traffic flow. Rate limiting wastes bandwidth and artificially suppresses demand and reduces investment in capacity! …or is the real end goal of a fabricated bandwidth constrained network to be used as an excuse for operators to increase prices?

Queue positioning techniques can help with latency control. These mechanisms can favor one traffic type over another at a congested location if implemented correctly. This differs from rate limited QoS in that it only comes into play when it is needed during periods of congestion and not mindlessly suppressing traffic rates all the time on underutilized links. Another key difference is that queue positioning schemes aren’t tied to individual users or individual services. The cost of the wire running into your home or business should not depend on what Internet services are running over it. The fact that you have already paid for Internet access is ignored as the telephony cartel wants you to pay twice so QoS can some how make some packets get to you “better”. In this case “better” is a secret revenue code for getting around artificially constructed p2p mitigation or traffic modulation impairments as well as artificial rate limited bandwidth constraints.

User controlled DiffServ (which by the way is a more benign form of QoS) can provide significant end user value by treating time sensitive traffic differently than other traffic at the usual point of constrained and sometimes congested edge capacity on the upload link leaving the home or business. I stress “user controlled” QoS because even queue positioning mechanisms can be abused by the operators who will continually look for ways to induce artificial bandwidth shortages thru perhaps shorter queues and/or higher drop probabilities if they can make a buck.

DiffServ marking under control of the end user can assist the positioning of packets at congested queues leaving or entering his home basically allows his network access link to run at a much higher utilization levels without impacting time sensitive traffic flows. This is a clear win - win, for the end user AND the operator because the operator can transmit more bits and makes more money with his infrastructure. But these useful queue positioning techniques are rejected by the telephony cartel because it doesn’t allow the operator to make more money by delivering fewer bits and charging premiums for QoS to alleviate individually rate limited end users.

All the QoS mechanisms being proposed today at numerous standards bodies is all about access and rate limiting with Telco cartel encouragement to tie it to individual users. At best it is a poor attempt to do some latency control by constantly restricting lower paying user traffic flows which lowers overall utilization. The net result is that overall, less bits are moved and therefore operator applied QoS not only discourages network use it also wastes capacity. QoS complexity, just like 3G is likely to become an unsustainable subsidized model marred by too high implementation and deployment costs.

Bandwidth Bandwagon

Bandwidth is usually enough! But, if you insist on using some form of QoS, perhaps the most intelligent thing to do today is to use the proposed Differentiated Service Code Points (DSCP in http://www.ietf.org/internet-drafts/draft-ietf-tsvwg-diffserv-service-classes-01.txt) BUT only under control of the end user for the following reasons:

- DiffServ is scalable,
- DiffServ is stateless,
- DiffServ works across ISPs
- DiffServ works equally well on access links and in the backbone,
- DiffServ has all the needed flexibility to work per hop or end to end,
- DiffServ works for both unicast and multicast traffic,
- DiffServ works for all types of IP traffic, not just voice.

Most importantly, DiffServ is an honest and proven technology. It was not invented by ITU shills to fill the role of a QoS telephony cartel revenue engine. IETF tradition requires trials of alternatives with running code, followed by testing for interoperability and finally a proposal for a standard. DiffServ should be the beginning and the end of the QoS debate.

Intellectual Property Restrictions

Maybe QoS complexity is really a hidden goal destined to further stifle Internet innovation because complexity is an ideal basis to drive Intellectual Property rights into the hands of a few companies serving their telephony masters. If looking at past behavior is any indicator, then QoS technology will surely degenerate into IPR as yet another means of market protection by setting the license costs for QoS technology at inordinate levels. As an example, GSM IPR is already high at a cost of 8-10%. For 3G it is over priced at 15-20%. ITU-T IPR behavior is simply designed to kill off any newcomers to the market who can’t bring their own essential IPR. QoS IPR will in turn help the Telephony cartel to remain in control. What could be better for the telephony cartel: invent controlling and rate restricting QoS technology that no end user wants or needs, force it into standards, control the QoS IPR, provide no end user value and in fact discourage network usage, and finally to add insult to injury extort money from the end users if they want adequate Internet service? The best outcome for the end users is for QoS to be defeated and not force the Internet to carry its burdens technically and - above all - economically.

Fundamentally, QoS as envisioned by the Telco cartel is about restricting and rate limiting the end user unless he pays extra and at the same time provides a means to charge content providers twice for the same bit hauling service. This also restricts capacity usage. Simple queue positioning techniques don’t care who is paying and enhances capacity usage. This puts the two schemes in proper perspective. Queue positioning mechanisms (such as user controlled DiffServ vs. carrier controlled rate limiting) are better because those of us with computer science or engineering backgrounds well accept the truth of the statement that work conserving scheduling disciplines, like using DiffServ for queue positioning, will outperform non-work conserving approaches such as rate and access limited QoS. In the end, they are both limited in what they can deliver and usage growth will eventually require additional bandwidth. But operator driven QoS is so much more complicated and expensive to deploy because extensive controls are needed end to end to maintain traffic preferences in the face of congestion, transport variations, equipment failures as well as mobility between network access points, to name a few.

Best effort is BEST!

The operator controlled QoS challenges and associated expense dramatically increase when you consider providing consistent QoS behavior, via policy, across multiple technologies, service providers, applications and devices. It is delusional to think that network service providers will share their policy information which they all view as a competitive advantage. This leads to the conclusion that best effort is fundamental for competitive behavior and makes end to end QoS an unreachable goal.

QoS is in fact open research, which means there is no accepted solution. There is also a clear danger that the cost of the QoS mechanisms needed, including their operation and management will far exceed what is required to provide ample non-QoS service, or simply bandwidth. The effort for QoS is like having multiple supervisors for every worker (just deploying any QoS scheme today is past the point where the law of diminishing returns negates the usefulness of control and therefore there is no ROI outside of extortion). The only view point without debate is that vendors pushing QoS love the complexity and know that it will drive additional product and services sales. Why not? There is always money for vendors to make by leveraging as well as contributing to increased network complexity.

Attacks on the anti-QoS perspective often center on the claim that the anti-QoS crowd want “infinite bandwidth”. This and similar statements by the QoS advocates are nonsense! Adding more bandwidth has always been more cost effective than QoS. Others also agree, please see http://www.dtc.umn.edu/~odlyzko/doc/internet.economics.pdf Think about this, if adding bandwidth wasn’t effective, then how come QoS isn’t widely deployed today? A look at QoS, past the hype, reveals numerous issues. As an Internet user, all I care about is that I have enough capacity so most things run nicely. Just give me my advertised bit rate and I am happy. I don’t need to be controlled or mitigated! In fact 600Kbs or so runs all the Internet stuff I want to do, which is a far cry from infinite bandwidth and I already pay for more than 600Kbs capacity today.

Investigation of Telephony cartel QoS propaganda reveals they provide no end user value and are part of the frankly appalling business models and inflexibility that incumbent Telco’s are forcing upon consumers and enterprise customers. In fact, QoS has lots of other known issues such as how it reduces reliability due to the excessive amount of state needed, it introduces enormous complexity, it has unknown impacts to security, it does not cooperate with admission control, it is even open research down to the meaning of the term “quality” applied to networking, vendor gear variations, lack of useful standards …etc. Basically QoS is not ready for prime time! Then why are some people trying to drive it into networking when it is not fully baked? This all begs the question: If QoS is so hard to do, reduces reliability, adds complexity, causes migration issues, etc. etc. etc, what could possibly be the motive of QoS advocates? So let’s think about this… It’s just a rate limiting and traffic modulation extortion attempt. It really doesn’t have to work, just limit rates and block competition, unless you pay a lot more of course!

Lies, Damn Lies and Statistics

In conclusion, QoS is all about telephony operators putting rate and access limiting mechanisms into the network to control the Internet, overcharge the end user with traffic surcharges, and restrict competition to simply line their pockets. I can’t blame the operators since monopoly protection results in insulated thinking; causes power induced greed, and clearly lowers business IQs which all lead to abuse (consider the parallels in operator arrogance shown by past “hush-a-phone” and now present municipal WiFi suppression activities). The really pathetic and distasteful part is that cartel (and increasingly Internet) vendors will support the QoS myth, co-opt standards and basically say, fabricate, and do “anything” for the anticipated future QoS product sales. But most troubling is; why won’t the government stand up for and protect the end user? Who by the way, vote! Maybe it’s time for true Internet advocates to use the powerful Internet tool they love and open web sites, blogs and information portals to bring this debate to the average American and especially to our government officials to clarify what an open Internet means to our freedom and to the global economy.

After reading the pre-conference interview I was very much looking forward to what Norman Lewis director of research for France Telecom had to say live at eTel.

I was not disappointed and as such have gone to the lengthy effort of providing a transcript of most of it. For those who want everything in 30 seconds, then I will summarise it before providing the transcript along with some comments that I made.

My personal summary: voice will become embedded in pretty much the same way that time has become embedded in everything (on wrist watches, cellphones etc). France Telecom will be offering an open platform via APIs for developers to combine the best of both worlds; grassroots innovation and a rock solid, telco scalable network underpinning. France telecom will collect and then presumably offer some kind of dip into customer data and metadata to provide underpinnings of exciting new services. FT will offer authentication, identity management functionality and customer care services to application builders. France Telecom sees the future as opening up innovation all the way to the edges and sharing their network assets with innovators. Applications of particular interest are powered by social networking and social software.

Norman was certainly in the top three of keynote speakers, both in terms of content, style and delivery. I think his speech was the most spot on I have heard come out of any telco, let alone an incumbent operator.

He began:

I would like to share with you today is a few ideas about the space we are in and how I see us going forward in the future…we are looking at three to five years ahead, we are looking at the intersection between the emerging and disruptive technologies and fundamental changes in customer behaviour and how they [customers] might interact with that technology in the future

Afterwhich he went on to highlight the disruption of the telco industry following the freeing of voice due to VoInternet:

I think it is true to say, that what has happened in the last years has been a bombshell [slide showed a nuclear bomb going off behind the city of Paris], and as you can see this is almost aimed at our HQ in Paris, I don’t have to tell you, this audience, why this has been so disruptive but I think the fundamental point is that when you have the commoditisation of Internet transport, where packets no longer care what they are on, what type of transmission, then essentially sound transmission just becomes one other capability of the Internet, and when that starts happening then a number of very important parameters fall away and new ones arise as you all know. The fundamental point is voice and audio now just becomes another application on the Internet. And that is incredibly exciting, as far as I am concerned, because it is like time, it is now liberated, it is not a stand alone application anymore. It is embedded in everything we do…Time has became intrinsic in everything. I think that is where voice is going in the future. I think that is truly revolution

He next turned his attention to the lack of innovation that telcos have exhibited since the birth of telephony:

I think that voice and innovation around voice has been very bloody awful. For the past 100-150 years I think telcos have throttled innovation around voice and voice has essentially been the same experience for the past 150 years…and we have that possibility of taking that application [voice]…and liberating it [voice] from that kind of stranglehold that I think telcos have had in the past… and now we can begin to do things we have never done before. …If you just look at the recent period with Ebay-Skype…voice is becoming something of an adjunct to other services and will open up new possibilities…I see this as a huge golden opportunity for immense innovation…What we [the telcos] are doing is re-arranging the deck chairs on the titanic. That is essentially what a lot of us are doing in our companies. The innovation landscape has changed…

He made some other interesting comments, including:

We [the telcos] are shackled by our business models. We are shackled by our legacy systems…business models are holding back our companies; I think it is pretty obvious that we earn billions on voice revenues. When you are, when the whole company is structured around certain revenues and certain business opportunities and when these come under threat, the immediate response is to try and hold onto what you have got, to try and stop the flow, to try and stop the disruption of values as you seen them in the past…they [telcos] don’t know which way to turn and as a consequence there is a real kind of stasis within the business…

He then took a swipe at 3G:

…even within that we still get it wrong…I can give you a number of examples, but the perfect example for me is 3G. In fact I was just talking to Peter Cochrane about this last night and who made the point that. I may not get these figures exactly right but the calculation is that if everyone of our customers, their families and their dogs had mobile phones and they used them for six hours a day for the next 31.7 years, we would get our money back on what we spent on 3G licenses…even on business models they sometimes screw up…why did they do it? It almost was that the financial markets were pushing things in such a way that if you were not in the process [3G] you were not seen as part of the future. In fact for being part of that process [3G] I think a lot of them in fact jeopardised their future for themselves

Back on the telco innovation front:

…every time the operator has done things [services] the customer has done something else. And as a consequence they have had to kind of follow the customer. This has been the history of telco. Think about WAP. You know that thing called WAP-crap… I remember the BT ads, surfing the web on your mobile phone, do you remember that? And when you got to it and of course when you got to it and it was like text and you pressed the button and you went away and made a cup of tea and came back and there were just a few words on the screen and you had to go down the menu, unbelievable. And what did customers do? They used SMS…they [telcos] did not even anticipate this [SMS] was going to be a service…

Personally I think he could have taken a swipe at a much more recent “BT innovation” that appears a fumbled mess, the BT Fusion which claims to be a Fixed Mobile Convergence (FMC) offering but appears completely ill thought out.

…so we screw up on our business models, continuously…the other thing that occurs, particularly in this world of convergence and integration between mobile and voice in particular is that one division cannibalises another. So we can have voice over IP, but that is going to screw up the revenue that we get over mobile. The mobile guys don’t want people to adopt voice over IP or fixed line telephony. They don’t want fixed lines in people’s homes. And at the same time we want to sell them DSL…So you have this kind of internal squabble going on, which I think creates a certain amount of paralysis as well.

He proceeded to provide a nice analogy of the ball and chain that legacy systems provide around operator’s ankles:

it’s like trying to drive only looking in the rear view mirror…the whole time you have to take into account how does that [non-legacy offerings] impact on our existing network, on our legacy systems, on our business models. And what that means essentially is that, the future is coming at you but what you are doing is looking behind you. You are looking everywhere except facing that future…And so as a consequence you try and you manoeuvre, you spend an enormous amount of energy, internally or whatever and essentially nothing ever changes

He provided another insight into the constraints facing telcos who try to innovate:

As a consequence of another great set of financial decisions that were made in the dot com boom, we had more debt than Russia. France Telecom had more debt than Russia. Now when you’re faced with a situation like that, you have a slight problem. Because you might be generating huge amounts of revenue, but your paying off huge debt… as a consequence the company has became more and more at the beck and call of the financial markets and you know, this is a reality that we all [telcos] face. And the problem with this is that it engendered a kind of short-termism that says it is not about the next three to five years but it’s about the next quarter. What are the figures, what are the returns going to be? And we saw the impact of this just a few weeks ago when we had to announce a slow down in the sales and give a profit warning. Our share price was badly impacted as a consequence…We are no longer willing to take the risks [of innovation] that we have done in the past. We are no longer willing to do the kinds of innovation that lead to the emergence of the Internet itself… I think this is a very significant problem. I want to share this problem with you, as I think it is a problem that is impacting everyone.

After selling the “Telco is Dead” pitch on the grounds of lack of innovation he swung things around. He pitched France Telecom as an emerging platform on which developers could work and in which collaboration with FT could take place in an open way leading to a potentially “beautiful relationship”. His pitch was that he wanted to break down the telco castle barriers, so that new innovation and thinking from outside the telco could begin flowing in. Again IMO he is bang on with the only long term solution to the telco is to decentralise innovation.

I was left wondering though after his acknowledgments that innovation surrounding VoInternet would be much greater than any traditional telco innovation, if he believed that an open France Telecom platform could be as competitive as the Internet to build on? I am hoping that he sees two future scenarios which should run concurrently. The first is VoInternet which can tap into the France Telecom network knowledge assets (CDRs, SCPs Etc.) and leverage such information (whom is on a POTS phone to whom, and derived social network structures Etc.) into extremely powerful Internet based applications. The other scenario is where developers can use the France Telecom NGN network as a platform for applications in which they will receive a cut of any generated revenue.

He laid the groundwork for the pitch he wanted to make by picking up and slightly shredding a previous speaker’s innovation:

I love some of the stuff we saw yesterday. But guys, you know, dialling up somewhere to get a message from Beavis and Butthead is not to be the pinnacle of innovation [the Voxeo speaker had given the associated B&B example that they had worked with MTV, the day before – actually I thought it was quite a smart app that they had worked). I think we have got to have higher standards; we’ve got to set our ambitions higher. There are enormous things we can do and have to do. But the problem is, if everybody is being tied into the short-termism, because if you are going to innovate, and you are going to get a product to market. You’re going to have to do it whether you like it or not, you’re going to have to do it thru an existing operator.

He immediately led this into a discussion of Web 2.0 business models:

What is the business model for Web 2.0 companies today? Correct me if I am wrong, and I am open to discussion on this. It seems to me that what most guys are doing, is they are coming up with a great idea, building an application and then getting bought by someone, getting bought by one of the big guys. Now that is monopoly capitalism guys, whether you like it or not, that is monopoly capitalism…what is the business model of Skype? I’m not sure about it, until now they have been acquired by eBay, now I see some real possibilities there…so whether you like it or not, you are going to be impacted by this mentality [short-termism].

After some preamble about both developers and France Telecom needing pray to a mightier force to overcome the short-termism together, he got to the pitch that he had been building up to, now that the audience was receptive he went on:

Now let me tell you why I think we might win [France Telecom]… It is a sincere attempt to engage you in an interesting dialogue about this, because I think there are synergies between what the aspiration that you have, and what we are trying to do. It can actually create a sweet spot for all of us…for me innovation is rarely about identifying problems our customers have got and trying to solve them. Real innovation is about social change. It is about adopting, it can be incremental, it can also be very disruptive. But if really had to begin with real social motivations, of why people are doing things. What kind of things that they really want to do… it is a social consequence that they [“digital children”] introduce technology into their lives in ways we do not quite fully understand… understanding customers [social] behaviour and motivations…that is the coal face as far as I am concerned…Are we going to develop Internet apps that really embed voice in everything we do, and fundamentally transform that whole experience. I think that is the question.

He went on to detail changes structurally at France Telecom and what they could offer developers:

France Telecom is moving towards an integrated strategy, full integration between the fixed, mobile, enterprise, home etcetra…we are in the process of completely rebranding our entire organisation as Orange and we will be the world’s first fully integrated operator. That really has some significant positives for the future. We have the networks, we have the scale, and we have the data. We will be able to amass data and metadata, about millions of customers around the world in ways I think everybody [people in conference] would dream of in terms of what you might be able to do with that data in the future…

He then spoke at length about customer support as a strong offering of telcos and gave examples of what do you do when your iPod has a problem and the fact there is no one to call and you need to trawl online forums looking for answers. He pitched that developers could utilise that support platform for their potential application customers:

We have quality of service… who do you call when it [your application] screws up… you can call us… We will solve the problems. We have that capability. We’ve been doing this for many many years. We have real people who know what to do. We have real technicians who can solve your problems there and then…I think that is a real value and that will become very much more important when we get into this Web 2.0…where we have web server environment and lots of services being cobbled together, mashed whatever, and something goes wrong somewhere. Someone’s going to have to sort that out. And we are the guys who can actually do that. We’ve got decades of experience of doing that kind of stuff. We can build for anything, we have that capability. We can authenticate and we can manage identity - absolutely critical in all of this [emerging telephony applications]…I think this is a key capability of ours in the future, customer care

He went on:

We [France Telecom] are the leaders in providing voice over IP and both France and in the UK, this is Wanadoo in the UK. We are one of the biggest providers in the UK of voice over IP today. Us, not BT, Us. In France the same thing. We’ve introduced some incredible pricing and packaging etcetera etcetera. Of course the real problem is you can get people to buy into the service but to get them to use it, is another question. And that is where this whole discussion comes a full circle. Because the way we are going to get people to use it, is by building compelling applications that embed voice, in other things that people are doing. And that is where I really want to talk about the future…by having this capability and more importantly understanding where we need to go in the future in terms of the sweet spot between providing that kind of infrastructure, providing that kind of capability with the ability to innovate that exists at the grass roots, that is where I think we need to come together and start working together and start building the next generation of applications that go far beyond the trivia that we see, ringtones…how stupid are people, you get a whole song for 99 cents why would people pay three pounds fifty [6.23USD] for 30 seconds of a song?

He finished off speaking of a project he called Octave that would offer the aforementioned APIs:

I would really like to make an appeal to you, that one of the things I am going to do for the future is a lot of the work that we are working on, and I can share some of this with you, we really want to make available to you guys… See by what we do in practice. This is a project that you will hear about in the very near future. It is called Octave. It is a personalisation platform. And coupled together with this we have another platform that we are developing around social software and [social] networking and we are going to make the APIs of this available and we are going to make available to you guys in particular, because we want you to be able to get access to the rich data sources that we are going to be able to orchestrate, and to be able to start building some of the applications and start experimenting with it. In particular around voice and one of the ideas that we really want to play with is using a social network and social software as a mechanism thru which you mediate and control context and your voice environment. So a caller would be routed according to where they stood in your social network and that would determine how you receive that call. And this would be a really easy system to manage as far you are concerned, in terms of the interfaces. Now I think that is really exciting when you start bringing all of those things together…we are going to make this available to the developer community and say guys, let’s work together. Here are the APIs, do with them what you will… it could be the beginning of a beautiful relationship.

Brough Turner typed up a nice little piece entitled Skype’s Part in Telecom Disruption. It covers the apparent fact (sorry for the oxymoron) that telecoms as we have known it for the past ~120 years has entered a state of upheaval due to three successive “disruptive innovations” acting on it. The first is mobile phones, the second is “rich communications” (listed are IM, presence, location information) and the third is speculative “eCommerce 2.0″ where telecoms becomes an adjunct of eCommerce. What he lists as “rich communications” is just the start of rich communications IMO.

By “rich” I would say that we mean the annexing and subsequent morphing of telephony beyond present day recognition; so that the very concept of a “call” becomes rather meaningless - more of a “session” at most. But even that is questionable since to initiate a session implies that people are disconnected in the first place. Personally I think further down the line that such an assumption could be wrong. Instead all people will be potentially connected all of the time with aggregators and filters splicing the self-broadcasts based on one’s interests, desires, and social group information as well as those of the “calling party”. They will help decide whom to keep a “warm social” constant thin bit stream with or whom to drop or whom to add. Instead of “calling” the user will wish to turn it to a “rich communications” stream for a period of time with one or more person/s. If that sounds far fetched then I will concede and say at the very least social groups will remain permanently connected in terms of bit streams flowing between them; presence and real-time location just to begin with.

As regards eCommerce 2.0 it is often said that markets are conversations! Or put another away by James Enck “voice is no longer about a battle for minutes, but rather a battle for consumers’ attention“. For a little more insight read Martin’s article Why Skype Needs Google which he wrote before Ebay bought Skype and also read that part about the Ancient Markets in the ClueTrain Manifesto (jump down to “storylines”). Even bricks n mortar commerce may be embracing VoIP to achieve more product sales.

The disruptive influence on telecoms that Brough Turner does not mention (actually it could be tied together with his “eCommerce 2.0″ theme) is the blending in of media into telephony (or the other way around depending on your perspective). I see no difference between telephony and other user generated and distributed content that is on the rise, such as blogs, podcasts, mashups, and independently produced content. These will invariably fuse with telephony.

The front page of the Economist the other day (15th September 2005) was “How the internet killed the phone business” (cover picture here). It is offered as premium content to subscribers at the Economist web site. But the bulk of the content can be found free in another article entitled “The Meaning of Free Speech” published in the same issue, see here. The latter article begins ‘The acquisition by eBay of Skype is a helpful reminder to the world’s trillion-dollar telecoms industry that all phone calls will eventually be free‘. Dialpad, another Skype-like firm, in June. AOL, Apple and others have similar products” and goes on to detail the Ebay acquisition of Skype that occurred earlier this month.

Quotes I note for interest/emphases are “even before VOIP makes 100% of telephone calls in the world completely free (which may take many years), it utterly ruins the pricing models of the telecoms industry. Factors such as the distance between the callers or the duration of a call, the key determinants of cost today, are simply irrelevant with VOIP”, “voice service is fast becoming a marketing freebie”. It quotes Cyrus Mewawalla, an analyst at Westhall Capital as stating

VOIP will destroy voice revenues faster than most analysts’ models predict…Voice will very rapidly cease to become a major revenue generator for all telecoms operators, fixed and mobile.

The article splits up operators into “unenlightened” and “enlightened” camps. China Telecom and Madison River Communications are placed in the former camp for blocking Skype and VONAGE, respectively. As is Clearwire and Vodafone Germany for reserving rights to block VoIP calls. Operators are placed in the “enlightened” camp if they offer a flat-rate calling plan or have launched their own VoIP service; Verizon’s VoiceWing and BT’s Broadband Voice are given as examples of operator offered VoIP. Three arguments are given for an operator choosing to embrace VoIP and thus be in the enlightened category; to compete head-on (as the article sees it) with the likes of Skype, to lower their costs dramatically and to “start offering the fun new services that VOIP makes possible and charging for them”. Exciting new services turn out to be video-conferencing, unified messaging and IPTV.

Video-conferencing can already be done with the likes of Skype, so I ignored that service. Unified messaging I find rather boring and thus I again ignored. That leaves IPTV which the article lists as the most exciting service. I think that says it all! To paraphrase “the most exciting service for telephone companies is delivering television”.

Switching to the fairly identical but shorter premium access only article that was front cover, quotes I note for interest/emphases are

the rise of Skype and other VOIP services means nothing more than the death of the traditional telephone business, established over a century ago…the ability to make free or almost-free calls over a fast internet connection fatally undermines the existing pricing model for telephony…that means not just the end of distance and time-based pricing - it also means the slow death of the trillion-dollar voice telephony market, as the marginal price of making calls heads inexorably downwards…it is no longer a question of whether VOIP will wipe out traditional telephony, but a question of how quickly it will do so…perhaps only five years away.

The articles states that mobile operators are the most vulnerable to Skype type services rather than fixed line operators - in particular 3G networks since they have a lot of license money to recoup, often provide unrestricted Internet access and have slow take up. This is plain wrong; Skype will hurt the fixed line operators first. 3G operators in my opinion have around three years fairly clear breathing space. This is because people are willing to pay for seamless mobility - which works! Skype type services will only be a big threat to 3G operators when 4G/Qualcomm/Flarion/WiMax/mesh networks emerge. Luckily Tomi Ahonen picks up on this mistake which saves me typing! For those without a subscription to the Economist, then at least for now both articles can be found here.