No Jitter is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

Telepresence Today and Tomorrow

WHAT IS TELEPRESENCE?

While the concept of videoconferencing-based "telepresence" has been around for over ten years, "telepresence" really entered the mainstream of enterprise consciousness at the end of 2005 when HP announced its Halo telepresence studios. Enterprise awareness then accelerated significantly a year later when Cisco entered the market promoting its CTS 3000 three-screen system and CTS 1000 single display system. Cisco's marketing and PR thrust was clearly intended to create a new market category dubbed "telepresence" and to then position Cisco as a leader in a new, exciting, and high growth segment. At the same time, Cisco's apparent strategy was to distance the company from videoconferencing, a solution set the company positioned as a 20th century failed technology. By any measure, it would appear that the company’s marketing strategy was overwhelmingly a winning one. Given Cisco's enormous clout with C-level executives, IT professionals, and the business press in general, the company was largely successful in grabbing attention (and market share) while at the same time creating confusion among the enterprise end user videoconferencing community.

Before the end of 2007, many of the established videoconferencing vendors felt compelled to react to the Cisco marketing juggernaut. Polycom, for example, re-labeled any of its systems capable of supporting 720p as "telepresence;" Tandberg defined a telepresence system as any system having a 50 inch screen after claiming that the company had shipped over 10,000 personal telepresence systems (previously known as executive systems) . Other vendors began to make claims about their "personal telepresence systems," many of which were high definition solutions based on laptops with webcams.

Wainhouse Research took a different approach by claiming that telepresence was just the high end of a videoconferencing spectrum and was not really defined by product attributes, but rather by the experience. In short, there is no such thing as a telepresence product, but there are telepresence solutions. We offer the following definition to the industry:

Telepresence is a videoconferencing experience that creates the illusion that the remote participants are in the same room with you.

HOW DOES TELEPRESENCE RELATE TO VIDEOCONFERENCING?

We believe that in order to provide the telepresence experience, a videoconferencing system must:

1) deliver life size images of remote meeting participants,
2) display remote images in the proper position and with the proper perspective, and
3) provide high quality audio and video signals.

To clarify our definition and the market segmentation, we have created the following diagram in order to properly position telepresence vs. videoconferencing vs. product segments.

Figure 1. Positioning telepresence versus videoconferencing

This diagram highlights the fact that there are multi-codec and single codec room systems and that multi-codec systems today represent the high end of the familiar group videoconferencing market. While most customers think of telepresence and multi-codec as the same thing, the vendors, particularly Cisco, include many single-codec systems (previously known as videoconferencing systems) as part of their "telepresence" offerings.

In fairness to the vendor community, nearly all of the multi-codec systems were originally designed to provide the "same room" illusion. Hence, they were sold as tightly-configured systems and installed in carefully controlled environments (sound, lighting, wall treatments, etc.) in order to provide a consistent, high-quality experience. In short, the original "multicodec" systems were sold to the telepresence buyers as fixed systems.

HOW HAS TELEPRESENCE EVOLVED IN THE PAST YEAR?

In the past year or so, we have seen four separate sets of developments impact the "telepresence" market.

Interoperability: It turns out that customers don't want to purchase and install $300,000 video systems in New York that cannot talk to the regional headquarters in Singapore or the sales office in Dallas--sites that cannot justify a full-fledged multi-codec system or sites where industry-standard room systems are already deployed. Furthermore, at least 50% of telepresence system owners have expressed the need to see individual contributors who may have only a software videoconferencing solution with a webcam attached.

Interoperability was not and is not a problem for Polycom and Tandberg, whose systems are based on H.323 codecs used in other systems. But Cisco's original position that interoperability with legacy systems was undesirable since it would break the "in-room" experience turned out to be not what the market wanted; the company then launched a full-fledged development program to enable such interoperability with its CTS codec systems, and it spent over $3 billion this year to buy interoperability and a complete standards-based product line through the Tandberg acquisition.

Flexibility: By the middle of 2009, the "telepresence" mantra began to shift as vendors began to react to customers' demands for "customized telepresence systems." Cisco actually led the charge here (although Teliris has long had a customizable solution and the LifeSize Conference system introduced in October 2007 is really a telepresence kit) by introducing customized telepresence systems in partnership with IVCi, one of the larger videoconferencing resellers and video managed services providers. The Cisco solution is based on CTS 3000 technology but allows IVCi to use the technology components to design to a customer's specific requirements or application. Tandberg followed suit shortly thereafter with its Telpresence T3 Custom Edition, as did Polycom with its ATX product. Ultimately IVCi formed its own Custom Telepresence Solutions Division to handle all three of these high end integrator products. Several other AV integrators and videoconferencing specialists have since been qualified by the manufacturers to sell, install, and service customizable telepresence solutions as well.

With these new telepresence custom solution offerings, customers can have systems where the size and number of screens and seats and tables and data display devices can be adjusted to accommodate widely varying needs. Few of these custom solutions are intended to give the same-room illusion that was the basis for Cisco's original telepresence product or market definition. These telepresence systems are really best described as multi-codec videoconferencing systems designed for AV integrators, similar to the way that traditional integrator codecs have been available for years.

The "telepresence" or multi-codec systems from the vendors are now much more flexible. And the vendor's claims about "in-room" illusions no longer make sense for the new customizable systems. At the same time, standard videoconferencing systems continue to evolve, with ever higher quality audio and video. We have found that with more integrators paying more attention to lighting, sound, and room layout, the standard videoconferencing experience is vastly improving. For those who insisted telepresence was a separate market, that claim is rapidly losing its meager foundation. Telepresence as a market segment is vanishing as the "in room illusion" and the quality of experience for telepresence and videoconferencing converge.

Figure 2. Telepresence is migrating away from the "in-room" illusion with time

Multi-codec Telepresence: We now come to an interesting dichotomy. While the early multi-codec installations were focused exclusively on meeting rooms and board rooms, more recently these systems have broken out into some very different applications based on the custom configurations now available. In early January, one of the authors attended a meeting in a large auditorium where a three-screen "telepresence system" on the stage provided exceptional audio/video during a videoconference. Based on huge, rear panel projection systems, the experience was extraordinary, but sitting in the audience, we did not feel at all like the remote participants were in the same room with us.

One fashion designer has used multi-codec systems to create a virtual runway to connect designers in Europe, the Far East, and North America in order to shorten the development time, design time, and fabric selection for clothing lines. These systems were customized to very specific needs and the installation bears little or no resemblance to the traditional enterprise meeting room. In a similar vein, multi-codec systems have been deployed in telemedicine applications to provide the highest quality AV experience without any wish to create an "in-room" experience.

While multi-screen doesn't necessarily mean telepresence, the reverse is also true. We have seen some single-screen videoconferencing systems deployed in ways that do indeed simulate an in-person, in-room meeting. With proper attention to lighting and sound, and with a sufficiently large screen at the right distance, a single-codec system can provide a telepresence experience for one or two participants on each end of the call.

Cisco's Acquisition of Tandberg: On April 19, 2010 Cisco closed its acquisition of Tandberg, a deal announced on October 1, 2009. While the impact of this merger on the overall videoconferencing market as well as Cisco's unfolding marketing campaign and product strategy remains to be seen, it is unlikely that Cisco will bad-mouth videoconferencing anymore or maintain that videoconferencing and telepresence are two separate domains, given that the company is sure to be the number one player in the visual collaboration market. We would not be at all surprised to see "telepresence" downplayed in the future as Cisco begins to emphasize its top-to-bottom video product line and its interoperability with legacy room systems as well with desktop systems being rolled out with unified communications environments.

WHERE IS TELEPRESENCE GOING?

The next 3-5 years are likely to see telepresence and videoconferencing developments in at least three areas of great interest to customers.

Interoperability:

The industry will continue its move to seamless interoperability between equipment from different vendors. Interop will progress in three dimensions:

1) Intelligent connectivity between single codec and multi-codec systems, whether from the same vendor or from different vendors;
2) Connectivity between multi-codec systems from different vendors; and
3) Integration with the unified communications environment.

In many companies, telepresence systems are a world to themselves, separate from all other communications mechanisms in the organization. That may not necessarily be bad, given that these rooms are almost always scheduled and managed. However, when those in a telepresence meeting need to temporarily bring in people from other locations with regular videoconferencing codes or work with an individual contributor who has only desktop video, more interoperability is required. We believe the percentage of companies who want the ability to mix telepresence with group and desktop video is quite high.

Cisco has integrated some WebEx capability into its Telepresence offering so that participants using WebEx can see those in a telepresence meeting and so that all participants, whether in the telepresence suite or using a PC, can share WebEx presentations back and forth. However, those in the telepresence room cannot see the video of the person on the PC unless that person manually connects through the Cisco Unified Video Conferencing bridge. Part of this decision to not allow PC video into the telepresence suite may be that the quality of the PC's video image would be poor--either terribly pixilated when it is blown up on a large screen or, if the image were not enlarged, it would appear too small on a large plasma display.

As HD cameras and compression capabilities become a reality at the PC level, look for more interaction between personal systems and telepresence suites. We also anticipate that as many companies roll out Microsoft OCS (and IBM's Sametime to a lesser extent) that desktop video will need to be able to be displayed on the telepresence screens. Interoperability between OCS desktop video and Cisco telepresence units would presently require a Polycom RMX or Tandberg VCS bridge connected into a Cisco bridge--which suggests lots of expensive hardware for a conceptually simple integration!

New Technologies:

Several new technologies will impact visual communications and videoconferencing in particular:

1) New cameras and image processing software are likely to provide wide angle, high resolution images that provide an enhanced conferencing experience without requiring the complexity and high bandwidth of multiple codec systems. Some of these devices may evolve from developments happening in the security market.

2) Multiple camera feeds along with new display technologies will provide more lifelike and more immersive experiences. 3-D immersive displays from companies like Telepresence Tech have become "show-stoppers" at many exhibitions; we believe the next generation of these displays will work their way into the enterprise market for both room and personal systems.

3) New compression algorithms such as scalable video coding (SVC) will enable high quality videoconferencing over lower-cost, lower bandwidth network connections, thereby making single and multi-codec systems more affordable to operate.

B2B Services:

Telepresence faces the same problem videoconferencing has faced throughout its existence, that being the "who can I call" dilemma. Early proprietary video solutions had limited utility because customers could have a videoconference only with others who had the same vendor's solution. In the early days, the way to connect videoconferencing codecs was through ISDN. ISDN may have been clunky, and sometimes hit or miss, but it did provide a uniform way to connect to other systems, and it provided a consistent dialing scheme--the phone number.

With enterprise telepresence, it is easy to connect systems from the same vendor, because they all run over enterprise IP networks. Thus, all an organization has to do is put in fat network pipes at every location where they want to have a telepresence solution. On the internal corporate network, the only real challenge is the expense of the network connectivity between sites (in addition to the cost of the telepresence suites to start with!).

But what happens when a company wants to have a telepresence meeting with a business partner, supplier, or vendor? Then things get a bit tricky. IP adds considerable complexity when one wants to use voice or video outside of the enterprise network. Companies intentionally put their video units behind network firewalls. Also, to conserve IP address and for some added security, companies add network address translation devices which give video units IP addresses that are not routable outside of the enterprise network. And, one of the biggest issues with telepresence is maintaining the quality of service required to have the "in room", immersive experience when communicating with people outside of the company using video.

The Rise of the Telepresence Exchange
One way to simplify inter-company telepresence communications is to have everyone you want to meet with subscribe to the same network provider and use the same codec. This is what AT&T, Verizon Business, BT, Telia Sonera, Telstra, Orange, and many other network service providers offer through a concept called a Cisco telepresence exchange. Conceptually, the idea is simple. If everyone is on the same network, then the provider of that network can assure that the necessary bandwidth is provisioned and that the quality of service on the network meets the requirements of the inter-company telepresence meeting. Participating companies would need to place a session border controller at the company/carrier network interface to provide security, or they could rely on the network provider's session border controller located in the network cloud. The network provider would also install a video bridge in its network operations center and tell those customers who want to join the meeting how to connect to the bridge.

In addition to the need for telepresence exchanges, telepresence directory services will be required to make the exchanges compelling. This reverts back to the "who can I call issue"--a telepresence directory is necessary to know who can be called using telepresence. Today, a telepresence exchange provider may offer a "directory" which contains a company's name and a contact person. If Company A wants to engage in a telepresence meeting with Company B, Company A must send an email to the contact person at Company B and arrange a mutually agreeable time for the video conference. Then Company A must contact the carrier and arrange for the ports on the video bridge as well as the dial in information. Then, both companies must dial into the bridge at the specified time. We believe that ultimately the telepresence directory must be on the network with enough smarts so that the telepresence exchange knows how to connect the systems together without all the manual intervention now required.

There are two challenges these Cisco telepresence exchange solutions face, however: 1) not all companies are on the same network, and 2) how do companies with non-Cisco telepresence units connect to those that do have Cisco telepresence solutions?

The Multi-Network Telepresence Exchange
To solve the first challenge, Cisco runs a multi-network telepresence exchange for its customers who are on different networks but who want to connect with one another. The concept is simple: Cisco buys network connectivity from a variety of carriers, runs these network connections into a data center, connects the networks up to a Cisco video bridge, and voila, those Cisco customers who ride a variety of networks can all join at Cisco's own multi-network telepresence exchange.

There are clearly some issues to be worked out with multi-carrier telepresence exchanges, including how the carrier gets paid for the traffic that crosses its network. Also, carrier companies will be sensitive to putting their customer lists in a public directory because this would allow another carrier to try to poach the customer. Other operational issues include things like who do you call if the call goes bad: if you are a customer of Carrier A meeting another company and you are both connected to a bridge located in Cisco's telepresence exchange, do you call your own provider or Cisco?

Another major issue with telepresence is how one connects telepresence video from Polycom, Tandberg, and LifeSize units with Cisco telepresence units. Bridging Cisco telepresence units is accomplished through the Cisco Unified Video Conferencing (CUVC) video bridge, which Cisco presently OEMs from Radvision. When a Polycom or Tandberg telepresence solution needs to connect to a Cisco telepresence unit, then the Polycom or Tandberg telepresence units must first connect to a Polycom RMX or a Tandberg VCS bridge, and these devices must then connect to the Cisco CUVC bridge.

Figure 3. Interconnecting Polycom and Tandberg telepresence units to Cisco telepresence suites

There are also other issues with integrating telepresence solutions from different vendors. For example, the algorithm for how and when the multiple screens should switch is not the same between vendors; hence, there is a need for some type of a telepresence interchange protocol .

(Cisco has submitted its Telepresence Interoperability Protocol (TIP) as a possible standard to the industry to connect multiple codec/multiple display systems together. There was no input from Polycom or Tandberg in this specification, however, so there is likely to be some spirited debate and some passage of time before such a protocol becomes standardized.)

Clearly Cisco would like to offload its multi-carrier telepresence exchange to someone else, and there is a movement afoot to establish other inter-carrier telepresence exchanges. Technically, this is doable--the problem lies in politics and in the dollars and cents.

Thus far, there has been little in the way of a compelling financial model for the various carriers to get together and provide a centralized, carrier neutral, telepresence exchange. Furthermore, a carrier supporting telepresence solutions and the associated network traffic does not want to lose customers to a competing carrier, so publishing a list of customers and routing addresses is not necessarily in the carrier's interest either. In addition, the carriers do not want to provide "telepresence peering points" with their neighbor carriers because there is no financial mechanism for doing so. The problem lies in the billing: does a telepresence-enabled company running over network A need to pay something to reach a partner, supplier, or customer who uses telepresence but is running on network B? If the amount of traffic were always symmetrical between network providers, there would likely be less of an issue, but a network provider does not want to get stuck with the possibility of carrying large amounts of telepresence traffic when that provider is not adequately compensated.

In spite of the challenges, there are multi-network telepresence exchanges that do exist or that are emerging. One of the leading B2B video exchange services is the Telepresence interexchange Network (TEN) run by Glowpoint. TEN allows anyone to connect to the video exchange and they can connect to anyone else who use the TEN network. The solution supports Cisco, Tandberg, and Polycom telepresence interconnect as well as interconnect between telepresence units and regular group systems, public video rooms, desktop video, and ISDN systems. TEN also provides a directory that lists who can be called (using E.164 numbers; E.164 is an ITU-T recommendation which defines the international public telecommunication numbering plan used in the PSTN and some other data networks. It also defines the format of telephone numbers. [Source: Wikipedia]).

An enterprise may inter-connect to TEN, with enough bandwidth to support multiple, simultaneous video sessions, for as low as $899 per system/month and can register external systems via the public Internet for $199 per system/month.

Figure 4. Glowpoint's Telepresence interexchange Network (TEN) interconnect service.

In addition to the Glowpoint TEN network, Tata Communications is creating a multi-carrier telepresence exchange designed for Cisco telepresence solutions. In Tata's solution, three geographically dispersed video network operations centers will provide redundant and highly available telepresence exchange capabilities for customers entering the exchange from a variety of networks. Tata will provision networks with significant bandwidth from a variety of carriers, thus enabling companies to use their existing network service provider to connect to Tata's telepresence exchange. Tata will bill by the minute and by the amount of bandwidth consumed.

Figure 5. Tata Communications' Global Telepresence Meeting Exchange configuration.

Unlike TEN, which supports interconnect between Cisco and non-Cisco telepresence units, Tata's solution does not natively support this interconnect. Consequently, when a Tata Global Meeting Exchange customer needs to interconnect with Polycom or Tandberg telepresence units, Tata uses the TEN network to do the interconnect between the non-Cisco telepresence endpoints and the Cisco endpoints.

CONCLUDING THOUGHTS

Despite all the hype and claims to the contrary, "telepresence" really is videoconferencing and always has been. The telepresence experience has taken videoconferencing back to its roots, with a focus on proper lighting, sound, room layout, etc. The new systems, whether single-codec or multi-codec, deliver the audio-video experience that people have always wanted from their videoconferencing systems.

There is still some distance to go with respect to interoperability and ease of use. Getting disparate systems to connect to one another is no small feat, as it requires expensive hardware and manual intervention. The telepresence exchanges that are emerging will help make interoperability easier, but at what additional cost to the end user company? Directory and call routing will need to be addressed to make these exchanges more compelling.

Our personal experience with telepresence is that it is a very compelling way to meet. If a company can afford the units, networks, and the services wrap, telepresence suites make wonderful meeting environments. We have also experienced properly designed conference rooms with high definition codecs as opposed to telepresence suites that are also very pleasant to meet in. Both kinds of meeting environments will continue to be used, although we suspect that most of the business will still go to the lower cost, high definition codecs as opposed to the telepresence suites. With regard to units sold, the present division in the market is 1% telepresence units to 85% group videoconferencing units (executive units make up the other 14%). This may explain why Cisco had to buy a videoconferencing company, since it was missing 85% of the market in terms of units and 75% of the market in terms of revenues.

Figure 6. The Q4 2009 group video conference market breakdown: units and revenues

ABOUT THE AUTHORS

Andrew W. Davis is a founding partner at Wainhouse Research and co-manager of the Video Communications Program (VCP). For more information about Andrew's video collaboration research work click here.

Brent Kelly is a senior analyst and partner at Wainhouse Research and co-manager of the Unified Collaboration Practice (UCP). For more information about Brent's recent work with hosted and managed UC service providers click here.