[LONG TERM PERSPECTIVES]   [CURRENT THOUGHTS]

Adapted from a Speech Delivered to the WINMEC forum
in March 2005.

How Telecom Providers Are Using Advanced Technologies Today

1. While the Market was Licking its Wounds, Telecom Engineers Were Busy
2. Rhetorical Question
3. IP and Broadband Have Similar Symbiotic Relationship as to Operating Systems and Microprocessors
4. Ramifications of a Converged Multi-service Network Driving IP-Based, Web-Centric Services and Applications Versus Separate Transport Networks for Distinct Services
5. Necessary and Sufficient Conditions for IP to Fully Displace Legacy Services
6. Can SIP/MPLS/Ethernet Do for IP What SS7/SONET/ATM Does for TDM?
7. Let's Examine Some of the Developments in the Different Areas to Achieve This Goal:
8. If Above Occurs, Will IP-Based Services Gain Commercial Acceptance?
9. Open Questions:

I.) While the Market was Licking its Wounds, Telecom Engineers Were Busy

While the stock market was licking its collective wounds over the meltdown in tech and telecom stocks over the past five years, network engineers were still busy working. In essence, we may be going from a period of "negative disarray" to "positive disarray", but disarray, nonetheless, which will present both opportunities and challenges to all players in the broad telecom/media/tech ecosystem. The result has been great advancements in standards and engineering to make IP networks closer to carrier-class quality, all of which occurred at a time of rapid deployment of broadband access over a variety of medium.

IP and Broadband will have a powerful symbiotic relationship which drives growth of one another much as has occurred with operating systems and microprocessors in the computing world. The consequence will be service features and applications that go well beyond existing offerings, providing a great deal of flexibility and value thanks to the virtues of IP and broadband. Furthermore, IP networks will retain the reliability and quality attributes of more costly and less robust legacy networks.

If true, we will see a proliferation of IP-based, web-centric services for both businesses and consumers. Businesses are likely to be the earlier adopters of packet-based broadband services given their need to enhance productivity and/or strategic position (i.e. Ethernet replacing frame relay and ATM for virtual LAN services). Consumer applications will likely lag due to both later buildout of networks (ubiquitous FTTH will come later than widespread packet aware, metro Ethernet networks in central business districts) and the fact that consumers are cognizant of overall household budgets and do not have as much "strategic" reason to be a leading edge user of IP-based services as do businesses.

The ubiquity of IP with billions (soon to be hundreds of billions) of addresses in cyberspace that can be reached anytime, anywhere versus legacy-based/location-connected devices will portend an array of new feature-rich services and applications which, over time, will diminish the importance of owning infrastructure. These applications will be able to be stored and retrieved anywhere in cyberspace so long as one has the appropriate IP address and can access the site where the application is hosted or where the information resides.

For example, corporate executives can access company-wide applications by connecting to an Internet link via a secured password which allows for greater flexibility and presumably enhanced productivity. Consumers could click on a web browser from anywhere they happen to be and have content delivered to a home recording device which has an IP address (an example of multi-casting). Moreover, the "industrial-strength" engineering of IP networks will make these services usable for both business and residential customers.

Furthermore, security takes on an entirely new and more critical meaning. In the old Bell System days, we had (and still do in the TDM world) monolithic networks and standards with managed devices (i.e. phones) which were physically connected to the network and features were embedded inside of physical elements of the network. In the new world order, we have a variety of IP networks (wired, wireless, etc.) with evolving standards with a myriad of "unmanaged" devices (i.e. PDAs, PCs, TVs, cellphones, iPods, etc.) accessing applications which may reside or be hosted on these networks or simply are accessed via one of these networks. The ability to hermetically seal a network is not possible like it used to be so advances in network security as we have seen in enterprises will be essential.

My view is that we are evolving from a TDM/SONET-based physical layer with an ATM data link and "connected" networks where services are very much tethered to the physical infrastructure to IP as the ubiquitous network layer with MPLS core providing "connection-like" carrier-class specs and where Ethernet is the universal layer two data link given its position as a true global standard within enterprises and its growing importance in carrier networks. Most importantly, IP-based applications will no longer be tethered to the network infrastructure, rather they will exist as hosted offerings on a multitude of networks which will change the value calculus away from pipes and onto application developers and further empower end-users.

The result will be a converged, multi-service network which means things are far more complex in terms of relationships between networks and services and where new applications reside versus the neatly compartmentalized world of the past. Applications will not be tethered to physical network infrastructure and will be the greatest source of value creation versus either the platform (i.e. IP backbone network) or the "power" (i.e. broadband access). IP will allow for various kinds of wired and wireless networks to carry applications which may be hosted on a network but is not part of one and wide scale deployment of broadband will allow such applications to proliferate but not, necessarily, be the domain of the owner of the physical broadband access pipe. Therefore, creativity at the application layer will derive more value than having a gold-plated transport network.

I believe, we will see IP over optical transport of packet-based services with an MPLS core providing carrier-class capabilities and Ethernet allowing for seamless transmission of services between locations across carrier networks without the need to breakdown and reconstruct packet streams. All of this will reach customers over a menu of broadband access choices ranging from BPL to WIMAX to high frequency fixed wireless to DSL and cable modems. Furthermore, these new services will, increasingly, be created at a different layer of the OSI stack than before when most services evolved out of layer one or layer two physical infrastructure. If true, this will forever alter the definition of "telecom" and will result in a kluge of networks and applications versus the clearly demarcated stovepipes of the past.

The IP/broadband relationship will result in the same kind of symbiotic relationship that the microprocessor and operating system enjoyed in the computing world for two decades with, in my mind, broadband representing the thing that enables or powers increasing IP-based applications much in the same way that advances in chips powered more features and functions on operating systems. The reverse was also true as operating systems became more feature-rich; the onus was on the microprocessors to keep up. The same is true in telephony, where new web-based applications such as multi-casting, storage and retrieval of files, IPTV, etc., are demanding higher bandwidth not just in core networks but also through the metro network and into the access fields. IP and broadband will drive each other's growth much the same way as operating systems and microprocessors have for one another.

The punch line is that IP, enabled by broadband delivery, will be the great disruptor in telecom in much the same way that distributed computing and power on the desktop was to the mainframe. The analog in telecom is that the old stovepipe structure will be dissolved with IP acting as the dissolvent and applications will be "up the OSI stack" and not, necessarily, tethered to the physical infrastructure. Value drivers will, increasingly, be from applications versus transport and those who are "bit haulers" will need to be smart about forming a nexus with new applications developers lest the owners of physical assets are relegated to being dumb pipes.

Back to top

II.) Rhetorical Question

Before adding some detail to the thoughts above, I think it is a good idea to ask a rhetorical question, namely "What or Who is a Telecom Provider?" A short history lesson is instructive in knowing how NOT to answer this question. One has to begin over 70 years ago with the "Telecom Act of 1934," (yes Virginia, the Act of 1996 was not the first one) which really set the definitions that, for the most part, still exist today. The 1934 piece of legislation envisaged a very clear separation of "church and state" (so to speak) by the way the authors laid out compartments of types of companies via what was known as "titles." For example, Title I was for ancillary services, Title II was for telegraph, Title III was for wireline and Title VI was for broadcast services (I am not sure what happened to Titles IV and V). Well, after all these years, wireless operators are regulated under Title II, landline telcos are Title III and cable is Title VI with so-called "information services" operating under Title I regulation which is the most lax. In other words, the companies today are nothing more than direct descendants of the regulatory stovepipes put in place 70 years ago.

The so-called landmark "Telecom Act of 1996" only tweaked what was then a 60-year-old piece of legislation, but, essentially, adhered to the traditional definitions of what constituted a carrier. This fatal flaw in the 1996 Act neglected to anticipate the arrival of an industrial strength IP network layer, the proliferation of devices which are unmanaged endpoints that are not connected to a monolithic network, and the transcending of applications across a myriad of protocols and networks. Hence, the structural change being brought about in this industry went right by the members in Congress and what was created by the Act only ended up causing havoc amongst both new and old carriers who were trying to play with somewhat new rules but thinking the overall structural framework of the industry was not materially changed. We will not even go into the changing regulatory rules which only served to confuse and frustrate new entrants…not to mention the economic framework of the Act being a disincentive to resale. This last point caused a huge buildout of network facilities which did not allow new players the luxury of building market share via resale - a la MCI circa 1973-1985 - but instead caused them to raise capital (largely debt) in order to fund capital expenditures which, among other things, led to the problems we all witnessed.

However, the real point of this discourse on the 1996 Act is to suggest that it should have either been created six years earlier or four years later. If the former had occurred, we may have actually had an orderly transition to a competitive local market with competition amongst players operating in a firmly established TDM world. More interestingly, if the Act had been passed four years hence, we would have at least had glimpses of developing feature and functionality of IP which could have drastically changed definitions of markets and, subsequently, could have resulted in differently defined but perhaps better equipped new players.

So the answer is…

A Telecom player is not simply "good old Ma Bell," despite efforts by some to resurrect her (it is kind of like going where the puck has been not where it is going - a la Gretzky) but now includes all sorts of other entities. While Verizon or Vodafone, obviously, still qualify, today, clearly the likes of a Comcast or a Con Edison qualify and, increasingly, companies such as IBM Solutions, Accenture, Cisco, Google, and scores of application/content developers are all taking their place in the vast telecom ecosystem. This is all coming about thanks to IP-based, web-enabled services delivered over an increasingly divergent variety of broadband access medium, hence…my view that IP is the great disruptor…the solvent which crumbles the 70-year-old stovepipes. For this to occur, the underlying network technology must still continue to evolve and these smart guys must develop applications and services that people will actually pay for using.

Back to top

III.) IP and Broadband Have Similar Symbiotic Relationship as to Operating Systems and Microprocessors

I view IP and broadband as akin to having the same type of symbiotic relationship as has existed between operating systems and microprocessors. In this analogy, I view IP as the operating system or software partner with broadband supplying the power or the "memory" that allows for more sophisticated applications. Just as we have evolved from simple spreadsheets and word processing on operating systems, the continued evolution of the IP network from a "best efforts" to a carrier-class network will allow for a multitude of connectionless, web-based services and hosted applications (i.e. multi-casting) which can be stored and delivered to or retrieved by users with IP addresses regardless of physical location. Similarly, the evolution from 286 chips to Pentium-class processors has given PCs far more memory and storage to allow for the enhancements to operating systems to be seamlessly delivered to computers.

Broadband access plays much the same role vis-à-vis IP as microprocessors have done with respect to operating systems. The so-called "Information Superhighway" had dirt road on- and off-ramps until we began to see accelerating deployment of broadband access. The continued rapid deployment of broadband access will enable the proliferation of IP-based, web-centric applications via more bandwidth (the analog to memory in PCs) directly to end-users. Thus, broadband access enables IP-based services much in the same way as faster processors enable more sophisticated operating systems and advances in IP will create more need for bandwidth much like advances in operating systems drove the development of faster processors with more memory.

This type of technology cross-pollination should result in a flood of new services with the result being a blurring…if not, a complete obliteration of the lines of demarcation between traditional telecom, media, and content players. However, if the only thing that comes out of all this is simply cheap transport, then it would have been a huge waste of time and money and brainpower. For example, VOIP is still largely marketed as "cheap minutes" which reminds me of MCI circa 1980 - "been there, done that." Rather, VOIP should be positioned as a platform for packet-based services such as hosted PBXs, unified messaging, simultaneous ring, conferencing services, and a full array of multi-cast capabilities. More importantly, a fully-engineered, ubiquitous IP network layer powered by widespread broadband access has the potential to take the concept of "on-demand" to an unforeseen level for a wide range of services that are content- or application-oriented. In the enterprise space, the advent of an IP/MPLS core will allow for scalable VPNs using the same IP network backbone as opposed to the current case where each customer VPN rides a separate IP network.

Clearly, the emergence of a converged, multi-service IP/MPLS network platform with Ethernet in metro networks and a variety of broadband access alternatives has great potential but also will be the source of the most significant structural change in the telecom industry. Unlike in the computing world where Microsoft and Intel were the two pillars of dominance for operating systems and microprocessors, respectively, in telecom there are no such players and the proliferation of IP and broadband will enable a multitude of players to develop value-added applications. Ironically, it is the big "bit-haulers," namely the few remaining large carriers who must figure out how to be in the critical path of value-creation other than simply having low-cost, high bandwidth transport.

Back to top

IV.) Ramifications of a Converged Multi-service Network Driving IP-Based, Web-Centric Services and Applications Versus Separate Transport Networks for Distinct Services

In no particular order, the following represent what, I believe, to be the major ramifications of the dawning of an IP/broadband world. The combination of an IP-network layer that is well beyond "best-efforts" and wider-scale deployment of broadband in access networks will result in the following:

1) Collapsing of the OSI stack, especially layers 4 through 7. In an ironic twist, voice goes from the bottom to the top of the stack. Specifically, in a TCP/IP world, IP is at the Network layer (3) with TCP at layer 4 (the Transport layer). In the legacy seven layer, OSI stack layers 4-7 are responsible for interoperability and "logical" connections between two streams of bits over physical connections (i.e. altering an email from a PC to be understood by a mainframe). I envision the Presentation and the Session layers (5 & 6) being subsumed within layers 7 and 3 or 4, respectively. As is the case today with TDM services, IP network services will depend on the data link layer (2) to deliver in this case packets (versus bits) from one hop to the next on the network. The difference between yesterday and tomorrow is that Layer 2 will, increasingly, be Ethernet as opposed to ATM and while Layer 2 services will continue to exist, increasingly, services will emanate from the IP network layer (hence my comment on voice going from the bottom - i.e. layer one circuit switched - to the top - VOIP at layer 7).

2) Applications move up the OSI stack and are separate from the physical layer. A web-centric on-demand game application will reside in layer 7 versus legacy services which evolved from the physical layers (i.e. private lines) and access to this hypothetical gaming application will be via "point and click" by the end-user and, unlike a CLEC needing a loop from a Bell, the layer 7 application will ride the customer's broadband access without ever needing to engage the underlying carrier of said access. Thus, the current regulatory ruling limiting the resale of carrier broadband access is irrelevant to a layer 7 application provider whose contact with the end-user is via the web. This is in contrast to a competitive Internet access carrier who needs to rent the DSL line from the Bell.

3) Horizontal versus vertical competition. By this, I mean that the competition for share of incremental value will be driven by competing layers versus competition between two stovepipes. As time passes, the question will be "at what layer of the OSI stack is most value created (i.e. applications versus network infrastructure)?" as opposed to "which stovepipe wins (i.e. cable versus telco)?" The multi-service IP core combined with a growing choice for broadband access means it will be difficult for any one entity to control end-to-end solutions, thus, "where on the value chain (i.e. OSI stack) will the most value accrue?" is the open question.

4) Bit carriers will become dumb pipes if all they provide is cheap transport. Owners of layer one and layer two network infrastructure will need to form a nexus with application and content developers perhaps in the form of guaranteed QOS and SLAs on packets going to a given application. Development of content is not, generally, found within the DNA of a carrier, so the aforementioned type of nexus could allow a carrier to participate in value-creation driven by new IP-based services.

5) Layer 2 will diminish in importance. ATM has been the legacy layer 2 data link for over a decade and services such as Frame Relay and other legacy packet services actually originate on layer 2. Going forward, layer 2 will deliver IP services but not be involved in service origination.

6) Ethernet will be the universal layer 2 protocol. IP, without question, will be the defacto network layer, and MPLS will enable QOS and SLAs for multi-service platforms to allow packet over optical solutions with SONET being ultimately replaced by packet aware networks.

Back to top

V.) Necessary and Sufficient Conditions for IP to Fully Displace Legacy Services

Despite the collective efforts of a lot of smart engineers over the past five years, IP-based services have yet to really make much headway in replacing legacy network offerings. The Internet was not designed to be a carrier-class network and is a "connectionless" network. This means there are no pre-ordained paths versus circuit switch nets with dedicated paths or an ATM type of packet switched network which creates virtual circuits that are erected and broken down. However, ATM carries with it a big overhead. IP, as a protocol that is becoming a network layer, is evolving from being a "best-efforts" network towards having carrier-class capabilities. This will allow for IP-based offerings to handle both new applications as well as enable the migration of existing legacy services to IP networks without sacrificing quality of service or reliability. Today, services such as voice, data, and video…not to mention music and games…still largely transverse separate networks, despite the fact that we are seeing some convergence at the device level. That said, the ultimate goal is to have multi-service network platforms based on IP with things like MPLS ensuring legacy-type network reliability.

For IP-based services to flourish there are both necessary and sufficient conditions which must be met. My view is that, in the ultimate irony, the necessary conditions which must be satisfied revolve around the ability of new converged, multi-service network platforms to, in fact, have many of the "reliability" underpinnings of legacy TDM/SONET/ATM networks so that IP-based, bandwidth consuming services can provide the necessary QOS/SLAs on a service by service level. Of course, having a network that is at once robust and capable of providing flexible, high-bandwidth services all with legacy-type guarantees is only half the battle. Thus, the sufficient condition for IP to flourish is that the new services are, in fact, commercially viable. IP-based, web-centric offerings must address the business needs of enterprises and/or the lifestyle needs of consumers in order for IP services to become pervasive.

Back to top

VI.) Can SIP/MPLS/Ethernet Do for IP What SS7/SONET/ATM Does for TDM?

The overriding question for engineers is "can they achieve in an IP world the same Quality of Service (QOS) and Service Level Agreements (SLAs) and overall notion of network "reliability" as we have come to expect in the legacy TDM world?" This entails perfecting "carrier-class" standards in things like SIP (Session Initiation Protocol), MPLS, Ethernet, and Packet aware networks. To do so would go a long way in establishing IP as the clear network platform of the future with packet-based services riding networks that offer dynamic, real-time reliability dependent on the requirements of a particular service. For example, instead of the 100% 1-to-1 protection of SONET which is expensive and unnecessary, IP/MPLS networks could label and tag packets by services and offer varying levels of protection depending on the service being carried by a given packet stream.

If IP networks can, simultaneously, deliver the flexibility and robustness they promise but, at the same time, offer service guarantees and security where required, this will go a long way towards quickening the migration from expensive and separate legacy networks to a true multi-service network platform. Of course, the market ramification of this will be revenue generation coming from creative applications versus simply reliable transport.

Back to top

VII.) Let's Examine Some of the Developments in the Different Areas to Achieve This Goal:

1) SIP is clearly a session control mechanism needed for flexibility in application development, supports real-time packet services, and allows set-up of sessions (i.e. services or applications) on IP. This is roughly an analog to SS7, otherwise known as "common channeling signaling" (or "out-of-band signaling") in the TDM world. SS7 is what allows an 800 number to be translated into a regular POTS number for example. SS7 also is vital in dynamic routing of calls. Similarly, SIP will set up sessions for packet-based services on IP network.

2) Ethernet is, arguably, the only true global protocol, being the instruction set (which is what a protocol is) that runs every LAN on the planet. I would argue that Ethernet is evolving from what I would call "Enterprise-class" to carrier-class as Ethernet works its way out of the four walls of office buildings into metro networks and potentially long haul networks (thanks to MPLS which will allow Ethernet to travel longer distances via so-called "tunneling" which forwards Ethernet packets through routed networks). Since, up until now, Ethernet was really enterprise-focused, the protocol was not designed for carrier-class network quality, not paying attention to bit-error rates, restoration of lost packets, round-trip delays, etc., and Ethernet speeds dovetailed with ports on equipment. Now, thanks to things such as RPR (resilient packet rings), Ethernet is being equipped with SLAs and the ability to modulate bandwidth in increments not solely tied to port speed of enterprise equipment. Given that corporate enterprises love Ethernet and would love to have multi-location services such as virtual local area networks (VLANs), the proliferation of Ethernet into carrier networks is simply a matter of how quickly it occurs. This will eliminate the requirement to mux/demux between Ethernet packets coming out of an office building and TDM circuits in a carrier network and then back to Ethernet at the destination site. Pervasive Ethernet in carrier networks will enable the deployment of cost-effective, feature-rich enterprise services on a multi-location basis. Thus, I believe Ethernet will replace ATM as the layer 2 data link since Ethernet is far less costly and is backward integrated into enterprise networks. Ethernet will serve as a delivery mechanism for packet-based, IP services with much less costly overhead than ATM.

3) The Sacred cow of SONET will slowly but surely fade away as traffic engineering at the MPLS layer allows for SONET-like capabilities but it could handle bandwidth consuming services. Just like Ethernet was not designed to be carrier-class, SONET was not designed to handle the transport of hundreds of terabits of packets which will eventually occur if FTTH really is widespread…not to mention IP-based services such as multi-casting and a proliferation of "on-demand" services. The idea of a stack of SONET rings the "height" of the Empire State Building, along with all the cross-connections that would be required, is unimaginable as a solution going forward for the transport of IP-based bandwidth consuming packet services.

4) MPLS, at the core of IP networks, is developing the traffic engineering which will allow legacy network-like quality and reliability but with the ability to more efficiently handle increasing requirements for bandwidth driven by IP-based packet applications. For example, MPLS can now offer SLA features such as bandwidth guarantees and management; MPLS will label packets which will give IP a "connection-like" quality, in essence, mimicking a PVC (private virtual circuit) with far less cost than ATM; packets will also be tagged (i.e. classified) over MPLS which will allow for SONET-type protection but on a dynamic basis, namely a voice packet will get 1:1 protection but a less time-sensitive packet will get 1:n protection with "n" dependent on the nature of the traffic within the packet. This reduces the requirement for costly overhead found in SONET for 1:1 protection for all traffic when, in reality, well over half the traffic on carrier networks do not need that level of protection. Advances in the QOS/SLA capabilities of MPLS will allow for packet over optical (versus packet over SONET or ATM) riding IP/MPLS core multi-service network platforms.

5) Packet aware networks will become pervasive in carrier metro networks eventually replacing most of the elements of old TDM infrastructure. Packet aware networks will easily interconnect with Ethernet interfaces which will be important because, as we mentioned earlier, Ethernet will become pervasive within metro networks. Thus, the ability to seamlessly interconnect carrier metro networks with enterprise networks at the packet level will allow for metro-wide offerings of such services as virtual LANs. In addition, packet aware metro networks will be equally important as FTTH and so-called "Triple Play" services are rolled out to households. I believe, switched Ethernet versus PONs (Passive Optical Networks) will be the architecture of choice in residential rollouts (more dedicated bandwidth per home and far cheaper than ATM delivery), and the consumer services being envisioned will be packetized, thus having packet aware metro networks will be important in carrying packets between homes and carrier installations.

The end result will be a multi-service IP/MPLS core backbone network supporting packet over optical services with packet aware metro networks and Ethernet being the universal data link with SIP controlling session flows. This will give the IP/broadband world legacy-like network quality and reliability while, at the same time, supporting in an efficient manner a myriad of IP-based, web-centric bandwidth consuming packet services.

Back to top

VIII.) If Above Occurs, Will IP-Based Services Gain Commercial Acceptance?

We discussed the "necessary" condition for IP to become a pervasive network off which an entirely new array of services are offered…that being getting the network elements ready for prime time. However, someone has to buy these services. Thus, the "sufficient" condition for IP to proliferate is the creation of applications and services that are doable because of the virtues of an IP/broadband network but also to satisfy either the commercial needs of enterprises or the lifestyle needs of consumers. The following are a few illustrative (but very real) examples of the types of services made possible in an IP world which will likely resonate with a large segment of either businesses or consumers.

One of the most widespread uses of MPLS will be the ability of carriers with MPLS core networks to build scalable VPNs using the same IP backbone. This is in contrast to today where each customer's VPN rides a separate IP backbone. MPLS will be able to ensure the isolation of each customer's traffic allowing the sharing of one IP backbone for multiple VPNs. Multi-casting services will be a huge set of IP-based applications and is truly a class of services that could not exist without IP. An example of such a service would be what is known as a "multimedia conference" (think of a teleconference on steroids) where audio, video, files, "whiteboards," etc. are shared to pre-approved IP addresses. The application will reside on a hosted platform on an IP/MPLS core network and is invariant to type of end device or access network so long as the device has the capability to handle the storage that is being streamed and the broadband access pipe has enough bandwidth.

Both of the above, along with things like virtual LANs which will be enabled by pervasive Ethernet at layer 2 and packet aware metro networks, are easily believable as types of service offerings corporate enterprises would find useful in helping their own productivity. For small-medium sized businesses, IP-based services off of VOIP platforms such as hosted PBXs, simultaneous ring, desktop-desktop conferencing and web-based user control of handset features will be very attractive, especially for those SMEs with tight budgets and lack of internal expertise.

On the consumer side, I believe, there is less clarity. Businesses will pay for services which, in turn, help them increase productivity or enhance sales, etc. However, most households have a limit on what they will spend broadly on entertainment/media/communications products. Thus, while IP-based services will certainly cannibalize legacy services in business (think Frame Relay), there could still be an overall expansion of the market if the take-up on the aforementioned services among business users is significant. In contrast, the consumer segment is more likely to be a zero-sum game with pricing being a much more important part of the value proposition, more so than it is for enterprises, which do not have non-price attributes to consider.

For consumers, the big buzz today is the so-called "Triple Play" of voice/video/data over either a telco or cable company network. However, these are just bundling of legacy services at a big discount. The virtue of IP coupled with broadband access for consumers will likely be found in applications such as IPTV which could allow a consumer to channel-surf on the Web, making on-demand video something that does not have to be part of a larger bundle which included ad-driven content. In general, IP will allow a slew of on-demand services such as on-demand ads. Clearly, streaming audio and video for music downloads or gaming will also be an example of IP-driven consumer applications. Also, the ability to seamlessly transfer content between devices (i.e. PCs, TVs, cameras, etc.) will require the ability to recognize IP addresses resident on different devices, and of course, last but not least is VOIP.

Ironically, it may indeed be the case that the coming of age of IP combined with the accelerating deployment of broadband may turn out to be the undoing of the so-called "consumer Triple Play" which was first envisioned more than five years ago in what was still a very solidly stovepipe-defined world. Now that IP can enable applications which transcend transport networks, the ability of the two big stovepipes serving the home (cable and telco) to dictate to the consumer what services they are offered may be undermined. IP will allow the applications developers to be the "brand" along with or perhaps instead of the network infrastructure provider. In fact, IP-based services are likely to be billed directly to consumers, bypassing either the Bell or the MSO, and such services may prove to be more appealing on a "when needed" basis than having to write a sizable monthly check for services one does not largely use. Of course, it is not surprising that if IP is indeed a disruptive force, it will be as ubiquitous in its disruption as it is in its virtue.

Back to top

IX.) Open Questions:

There remains a slew of open questions that will be answered over the course of time. The following are a few to ponder.

1) Broadband access is what powers IP. Will broadband penetration double from current levels over the foreseeable future? Currently, only 5% of broadband homes have VOIP, can that grow and can other IP-based services achieve better take rates with consumers who have broadband connections?

2) The virtue of IP is the ability for applications and services to be hosted or reside anywhere given the ubiquity of IP addresses and the pervasiveness of IP networks. The commercial question will be where will new, value-added applications reside? Will they be on carrier networks, within enterprise networks, in a router in the basement of an office building, or as part of a systems integration solution? This will go a long way in determining the relative share of value-creation amongst the various players in the telecom ecosystem.

3) Security. As we mentioned at the outset, we are evolving from a monolithic network with managed endpoints to an environment with unmanaged endpoints with packets carrying services transversing different networks with varying protocols. Will intrusion protection be able to scale for wide area networks as it is beginning to do for enterprise software and operating systems?

4) Role of wireless. We have only barely mentioned wireless in this discussion because wireless networks are really access networks…be it mobile or fixed…with the laws of physics putting limitations on the amount of bandwidth that can be carried on RF spectrum. Having said this, a real question is can WIMAX with issues surrounding unlicensed spectrum, lack of standards, no mobility, and still high cost points, become a true broadband alternative? As we speak, WIFI is a huge success from a user perspective (I am not sure any money has been made) and not just in places like Starbucks. Enterprises love having internal WIFI access points as a perfect way to take advantage of the virtue of IP and Ethernet so that a worker can travel with their laptop and still be able to be contacted anywhere in a corporate environment (SIP phones would do the same thing for a phone number), but WIFI is range-limited so other fixed wireless solutions need to prove out both economically and capacity-wise. As far as mobile wireless is concerned, a real question is the tradeoffs between more bells and whistles versus the ergonomics of the phone and the power consumption?

5) Which market segments will most benefit from rollout of new IP-based applications? It seems, at this point, an almost "barbell" shaped market is forming with either Fortune 2000 enterprises or residential homes being targeted by the various large suppliers or carriers for rollout of new service offerings. The vast middle, namely small and medium businesses, has so far been largely neglected. Clearly, the two big recent telecom mergers are largely driven by the desire of the acquirers to gain traction in the large enterprise space. Meanwhile, the FTTH rollouts and overall "triple play" offerings by both cable and telcos are targeting the residential user. Thus, a question remains as to the degree to which IP-based, web-centric solutions will be actively marketed to the SME segment? The irony is that SMEs, who do not have the in-house budgets not expertise of large enterprises, are prime candidates for hosted IP applications. Perhaps this time around there will, in fact, be a place for niche carriers serving this market segment.

6) Public policy - what will it be? Our view is the less the better, let the marketplace decide.

7) Finally, at the end of the day what does all this mean to the industry revenues and profits and structure? What will the net impact over the course of the next 5-10 years of the rollout of new IP-based services be? Will the cannibalization of some legacy services be offset by growth in new services? Will the industry be able to respond appropriately to a revenue model that is moving away from transport of bits to usage of applications? Will the industry continue to have just a few behemoth carriers or will IP truly be the solvent which results in a much more fragmented industry structure with players and brands in segments that one would not necessarily envision in these segments today?

Back to top