Last week, Verizon announced its much-rumored intention of pursuing 4G wireless Long Term Evolution (LTE) standards instead of Qualcomm's Ultra Mobile Broadband (UMB) technology.

With Sprint pursuing WiMAX with its Xohm announcement, two of Qualcomm's major North American accounts have indicated alternate plans instead of continuing down the Qualcomm CDMA evolution path toward 4G.

If this trend continues worldwide, Qualcomm will need to bank on revenues elsewhere, perhaps by embracing LTE as well as WiMAX, although Qualcomm has resisted WiMAX previously. With newer software-defined radio chipsets emerging where UMTS and WiMAX offerings exist from vendors like picoChip, perhaps Qualcomm can expand this toward LTE and WiMAX. As dual-mode devices proliferate, along with emerging femtocells that combine macrocellular with broadband access, more flexible solutions will be required.

With UMB, LTE and WiMAX being based on Orthogonal Frequency Division Multiple Access (OFDMA), it also remains to be seen how valid and enforceable Qualcomm's patents are in this regard, especially via its Flarion acquisition, which expanded Qualcomm's OFDM/OFDMA intellectual property portfolio. This may also be a source of further revenue, although more analysis will be needed.

Submit a Comment | Permalink

Teliris, with corporate offices in New York City and London, just announced the release of its Telepresence Gateway - a product that enables Teliris VirtuaLive telepresence customers to connect to competing standard telepresence systems, as well as legacy videoconferencing, audio and desktop conferencing systems, and Web-enabled devices such as mobile phones and PDAs.

The release of the Telepresence Gateway will initially support interoperability with Polycom RPX and Tandberg Experia telepresence solutions, with interoperability support for Cisco TelePresence and HP Halo down the road (see "Comparing Videoconferencing Telepresence Systems"). This is a significant announcement for the market because telepresence solutions from high-visibility vendors like Cisco and HP are just starting to address interoperability with legacy videoconferencing systems, which Teliris has been doing for the past five years.

The Teliris Telepresence Gateway rolls in the legacy interoperability function, along with support for mobile devices, while also providing what the other telepresence vendors don't do yet - interoperability with other vendors' telepresence solutions (see "The Gartner View on Enterprise Video").

1 Comment | Submit a Comment | Permalink

Some of the emerging telepresence systems can create quite a rush of interest and acceptance from users. Enthusiasts of telepresence describe it as creating the impression or illusion of being present with other people over long distances.

Telepresence uses new high-end videoconference systems to facilitate virtual meetings, which users say are almost like being there. The new high-end systems use broadcast quality cameras and typically a set of three 50-inch plasma screens in specially designed rooms. A telepresence room is usually limited to six users sitting along one side of a conference table with two seats facing each screen. A lifelike effect is achieved by placing cameras behind the screens, so live participants are shown staring directly into the eyes of the remote person.

These are not the conventional videoconferencing systems, plagued by jerky video and speech that are in no way synchronized with speakers' lip movements. This is a whole different level of system that should probably not be compared with videoconferencing and its historical failures. I avoid videoconferences, but I would use this. It's quite astonishing how well the illusion works, at least in demos. You can completely forget the technology is there, and you have your meeting just as if you were face to face. Telepresence equipment makers and users say the experience is quite different from normal videoconferencing. Some of my colleagues are adamant that it is not useful to think of telepresence as just "good quality" videoconferencing.

Its impact on users is far more substantial than that. It should be considered not only as a travel displacement tool but also as an alternative to the corporate jet.

Many businesses (especially large enterprises) have videoconference rooms, but in my experience as an operator of these facilities, their poor quality meant they averaged only about 10% utilization a month. Early telepresence installations seem to have much higher utilization rates, with many booked solidly throughout the day.

The real enthusiasts seem to be those who have avoided frequent (and often exhausting) intercontinental travel.

Telepresence systems cost about $500,000 for each location and may add $10,000 to $20,000 a month for the extra 2 Mbps to 45 Mbps of bandwidth required. The three-screen 1080p HD versions require 12 Mbps. This puts telepresence costs at 10 to 100 times more than conventional videoconferencing.

Great news for carriers!

There will be carrier opportunities here. The rooms are expensive, and it consumes a lot of bandwidth, so we're likely to see carriers offer facilities in major cities rather than having a lot of end users buy this stuff.

Is it worth it? Blog me and tell me what you think.

So, who are the players?

HP sells a system called Halo, with each equipped room priced around $400,000. Market leader in conventional videoconferencing Polycom has telepresence solutions from $200,000 to $500,000 a room, compared to its videoconferencing, which sells for $10,000 to $80,000 a room. Tandberg, Teliris, Telanetix, DVE and TeleSuite (now Destiny Conferencing) also play in telepresence. Cisco has been working on algorithms to compress the bandwidth. It claims 2 Mbps per two user sessions with 1080p HD. CallManager runs the event. Cisco seems to plan on being involved in consumer videoconferencing - via its set-top boxes (Scientific Atlanta), its home wireless technology (Linksys) and through the growth of IPTV.

I suspect that telepresence needs to be treated as a distinct category. Calling it high-resolution videoconferencing probably misses the point. It can reasonably be called a virtual meeting environment that happens to be enabled by high-end video and audio. Despite the price and the fact that sales in the last year or so are only around 100 globally, there are applications and a market for it.

Most enterprises won't buy these systems; they will lease time in conference centers, hosted venues and hotels. Some large (especially global) enterprises will buy these systems. Smaller companies will lease time because they can't afford them (or don't want to spend the money), and larger companies may lease time to try them out for a while until they can justify and afford the expense.

Gartner expects that until 2010, the use of video in corporate environments will bifurcate into very low-end and very high-end markets, putting tremendous pressure on the current midrange room-based systems. Use of low-end desktop USB-based webcams will grow because they are so cheap. High-end "executive-jet-class" systems will grow because the quality is so good. The current model of fairly expensive, hard-to-operate, medium-quality conference-room-based systems will slowly die away in favor of either of these two emerging models.

What's your view?

Submit a Comment | Permalink

PowerDsine, based in Israel, and a leader in the development of a technology called Power-over-Ethernet (PoE), said it will be acquired by Microsemi Corporation. Microsemi is a semiconductor manufacturer located in Irvine, CA. The PoE technology, also called Power-over-LAN, allows network appliances to receive power over existing LAN cabling. It is an international standard, known as IEEE802.3af, that was ratified in June 2003.

PowerDsine delivers its PoE chipset solution through networking vendors such as 3Com, Nortel and HP, which then embed the technology into their switches. Cisco –which is the leading switch vendor- is not one of their clients, but PowerDsine is a member of its Technology Developer Partner Program,. PowerDsine has also tested and certified various PoE-compliant network devices from numerous vendors to ensure conformance with the IEEE 802.3af standard. As many switches do not support 802.3af, PowerDsine also sells its own branded PoE midspans, which are a standalone device sitting between the LAN switch and the powered device, but sales have been slightly disappointing.

There are two main applications at the moment driving the purchase of PoE-enabled devices: VoIP and WLAN. There are also numerous other applications, ranging from very niche ones (Gibson released a PoE guitar) to some with a reasonable potential market (security cameras and automation), so we expect sales of PoE switches to increase 4-fold by 2009, to reach over 100 million ports. PowerDsine has been pushing for the use of PoE in as many different types of devices as possible. It recently added Wireless Intrusion Prevention System to the list, through its partnership with Airtight Networks. PowerDsine is also working on the development of a technology called PoE+, which will enable devices like laptops to be also powered, as they require about 30-60W, which is substantially higher than the maximum power of 15.4W that the current standard allows for.

With this acquisition, Microsemi accomplishes a number of things, including:
• Expanding the portfolio of product offerings to include a broad range of Power over Ethernet solutions
• Enhancing its analog and mixed signal design capabilities
• Getting access to some new, unfamiliar markets, such as communication and networking, as well as direct access to the pool of PowerDsine customers.
• Expanding its marketing and sales presence into new geographies
• Growing its customer base
• Raising the volume buying power with suppliers.

Finally, both companies will be able to share their power management knowledge, and expand their analog/mixed signal portfolio.

Submit a Comment | Permalink

On 11 September Free (a subsidiary of Iliad) sent out four press releases concerning a decision to roll out fiber-to-the-home (FTTH) throughout Paris by the end of 2012 - covering more than 10 million people, corresponding to more than 4 million potential network connections. In a follow-up press release on 12 September, Draka Comteq was announced as the supplier to the first phase of the project, under which 2.1 million people during the next 24 months are to be offered FTTH services.

While the move from Free is bold in terms of size alone, it is equally bold in terms of the associated business strategy. To achieve a quick ramp-up of services following the aggressive deployment, Iliad has announced the following:

• A €29.99/month service package (which includes provisioning of an optical terminal) will be offering 50-Mbps broadband access, unlimited phone calls to fixed lines in France and to some international destinations, and HDTV content.
• Free access (deposit required) to a telephone line without telephone subscription, emergency calls and calls to social services, low-speed Internet access (e-mail and Web access) as well as antenna service with access to free TNT channels in digital quality.
• Free has announced that, once the network is up and running, talks will be held with all interested operators with a view to offer FTTH leasing agreements. Free states that "the leasing price will enable operators to replicate Free's commercial offerings."

With the above offerings Free is taking an approach that ups the ante in the French broadband market in several ways. The nearly ubiquitous FTTH service addresses the advanced part of the broadband segment, and will in the longer run be able to scale bandwidths beyond what non-fiber infrastructures can reach. The free service and the wholesale offers make the move politically correct in terms of bridging the digital divide and enabling fair competition among service providers. But the free service and wholesale offers also have merits in a more narrow business context: The free service will make Free attractive to the lower end of the broadband market, thus stimulating the migration from legacy infrastructure to Free's FTTH infrastructure, and wholesale revenues will benefit the infrastructure part of the business case in terms of lower payback time and risk mitigation against customer churn.

It will be interesting to follow Free's execution on the announced strategy as well as the reaction from competitors in general and France Telecom in particular. If Free is able to execute according to the announced objectives, the company will effectively become "the FTTH incumbent" in Paris. The real question therefore is: Will France Telecom try to compete with Free in the lower parts of the value chain by speeding up its own infrastructure investments, or will France Telecom decide to put emphasis on differentiating its service offerings and simply become a third-party operator over Iliad's infrastructure? Incumbent carriers have talked about climbing the value chain for more than a decade - this will be an interesting litmus test for how far they are willing to go.

For more Gartner research on FTTH deployments, see for the following:
• "A Business Model for Next-Generation Broadband Access (February 2006 Update),"
• "Funding FTTx and Expanding Your Marketplace,"
• "Danish Utilities' Grand FTTH Plans Could Set a Trend,"
• "Why Governments Should Care About Fiber-to-the-Home,"
• "French ISP Free Must Move Fast to Secure Triple-Play Success."

Submit a Comment | Permalink

A recent voice-over-IP (VoIP) study conducted by Brix Networks confirms what most of us already know: VoIP quality is erratic. The study asserts that quality has gone down over the past couple of years, but our reading of the data shows otherwise. A cut of the most recent data shows that VoIP quality is improving!

Measuring quality on the Internet is a horrendously difficult proposition. There are dozens of backbones, hundreds of millions of endpoints, and tens of millions of people running VoIP. To come to any meaningful conclusion, you need to take a lot of measurements from a lot of different endpoints. You need to be able to test real voice calls, including peer-to-peer calls. You need to be able to capture the users' experiences with the actual software and service they use, such as Skype.

But the Brix test falls short: It measures people who have an interest in measuring performance (read: those that may be experiencing problems). The test emulates an SIP voice stream, but is not actually a call. It derives quality based on packet metrics, using a technique called the E-Model - not the more robust (and processor-intensive) PESQ algorithm used in some of the Brix testers. It does not measure calls between peers, but to a few centralized hubs. Measurements are actually collected at only seven points - nowhere near enough to make a universal statement about the Internet as a whole.

Methodology aside, the measurements over the most recent six months show a marked improvement in quality. See here - a trend line for the first half of 2006 would move upward, not downward. To further my point, the largest volume of samples has come in the most recent three months, which shows an even sharper increase as the service has become popular. Getting worse? I don't think so.

But there is still something to be learned from the study. Voice quality is dependent on many things - especially latency, jitter, loss, codec choice, endpoint tuning, and - less frequently than you might think - bandwidth constraints. A high-quality network is essential, and many of the variances in the Brix study are based on the quality of the pipes underneath it all.

Some providers are better than others, and bottlenecks appear in many points of the network. Some providers oversubscribe their access circuits by 10 times or more - which creates congestion getting to the backbone. Others may have limited and underprovisioned peering relationships with other Internet providers - this is congestion between backbones. A few suspect the providers themselves of impairing or even blocking competitive voice services - congestion inside the backbone.

There is good VoIP and there is bad VoIP - and there is even some good VoIP being done on the Internet. But not all VoIP strategies are the same, and it's important to distinguish between approaches. There are roll-your-own approaches based on corporate telephony platforms, used for the majority of the enterprise. There are broadband IP telephony services like Vonage, which may be good for teleworkers. There are personal internet telephony systems like Skype that may be more appropriate for mobile workers. If you're building it yourself on your own network, doing a detailed assessment is critical, and measurement solutions like those from Brix can help. See "Readying Your Network for VoIP".

So VoIP should be nobody's only means of communication. But it is a part of the bigger portfolio, including traditional landlines, cell phones, messaging systems or corporate-provided VoIP solutions. Like any portfolio, you want to spread the risk around. Selectively using the Internet for a corporate environment is smart. See "Is the Public Internet the New Network of Choice for Enterprises?". Just don't bet your business on it.

Submit a Comment | Permalink

ABN Amro has announced that it is introducing voice verification for its telephone banking customers in the Netherlands. Voice verification compares certain biometric attributes of the caller's voice with a record captured during a prior enrollment process, to verify or challenge the claimed identity of that caller.

One of the difficulties faced by companies that wish to provide customers automated telephone access to services, including confidential information, is to provide a secure method of identifying and authenticating the caller cost-effectively and without making it unduly difficult for authorized callers. Typically, most companies shy away from automating this part of a caller interaction, which limits the scope of services that can be automated. Most companies have relied on call center agents to identify callers. Not only is this process expensive, but it is subject to security risks in itself. Voice verification has promised to help overcome this difficulty, allowing callers to interact directly with a system to identify and authenticate themselves, without the need for call center agents.

However, it has been slow to be adopted due to difficulties with accuracy and motivating customers to enroll on the system, and there are few long-term, large-scale consumer-facing applications. With continued concerns on the part of companies with regard to striking a balance between security and allowing customers easy access to services, vendors have continued to invest in devising solutions. This announcement by ABN Amro highlights that progress is being made. It will eventually lead to voice verification being widely used to provide a secure method of providing customers telephone access to systems containing confidential information, such as is held in bank accounts. Companies should research the technology and match solutions carefully to their applications. We would generally recommend starting with internal employee-facing applications, such as password reset, before moving on to customer applications, and even then, piloting with your own employees, who are also customers of the services, before scaling out to a larger customer base. (See "Voice Verification Makes Big Strides, But It's Still Risky" and "State of the Art for Online Consumer Authentication.")

Submit a Comment | Permalink

The WiMAX market in the U.S. received another boost of activity as Intel announced its US$900 million investment into Clearwire, a broadband wireless service provider headed by Craig McCaw. As part of the announcement, Motorola will acquire Clearwire's wholly owned subsidiary NextNet Wireless, an equipment company for proprietary broadband wireless technology. Both Intel and Motorola stated their desire to expand broadband mobile data and content-rich Internet services throughout the U.S. Clearwire had an IPO filing in the beginning of the year, trying to raise around US$400 million, but now is off the hook to demonstrate that its business case really does work. Clearwire owns licensed spectrum in the 2.495GHz to 2.690GHz band. The company has the second largest spectrum position in this band after Sprint Nextel. It also owns spectrum in international markets in the 3.5GHz band and has started services in Denmark and Ireland. Owning NextNet meant that Clearwire was able to optimize the technology performance for its own purposes - but at a cost. The Motorola acquisition of NextNet puts Clearwire in a precarious position. On the one hand, it needs to continue its service rollout and does not need any interruption in the supply of technology; on the other hand, becoming 802.16 compliant will offer scalability and cost efficiencies. Motorola clearly gains the intellectual property of NextNet's technology.

For Motorola and Intel, the strategic advantage is that they cut off a potential irritation in the push for WiMAX acceptance in the U.S. Clearwire had always contended it would look at WiMAX when it made sense but would pursue its proprietary track. A convincing success story here may have raised some eyebrows in the WiMAX community about the significance of the business case for WiMAX (see "How Network Service Providers Are Approaching WiMAX"). However, the U.S. service providers don't have access to 3.5GHz spectrum, so at the moment, they also have no access to WiMAX-certified products. In theory, Clearwire had a justification for its model. That would have changed, though, because the WiMAX Forum will start certification for the 2.5GHz range as well.

The entire discussion about bringing mobility into the broadband wireless market is overhyped. Gartner's position is that 802.16e 2005 won't be able to provide true mobility in scale until the end of 2008 (see "Market Forecast: WiMAX, Worldwide, 2005-2010"). Semi-mobile applications, however, will be a market opportunity, connecting metro areas or MANs. Clearwire offers a great testbed for Intel, because it owns enough spectrum to cover more than 90 million people in the U.S. We conclude the U.S. broadband wireless market is moving in the right direction. Still not convinced? BellSouth announced in a separate statement that it will expand its broadband service to five new markets, with pre-certified technology in the 2.3GHz spectrum. It is also trialing the 802.16e 2005 solution by Alcatel.

Submit a Comment | Permalink

Intel announced on 28 June 2006 the sale of its communications and application processor business unit, which was focusing on providing the base technology for end-user handheld devices. For a moment, it seemed that its interest in WiMAX was at stake. However, the company confirmed that it is still committed to the WiMAX process. The question occurs: How dependent is WiMAX on Intel? Would WiMAX lose credibility or momentum if Intel opted out?

Gartner has always identified the industry discussion about mobile WiMAX as hype and has instead redirected and emphasized the benefit of WiMAX for broadband wireless DSL extension, SMB fast data access and private networking. For 802.16e 2005, Gartner did not expect full mobility to be deployed before 2008, but rather partial mobility for metropolitan networks. Therefore, we believe that the discussion about mobile WiMAX needs to focus on what the real applications will be, rather than what the technology in theory can deliver.

Intel is a strong contributing member of the WiMAX Forum. However, the WiMAX Forum has successfully implemented the certification process - with vendor products already obtaining the certification label. Simultaneously, the organization is preparing the road map for 802.16e 2005. So one could argue that there are dynamics and credibility created by the organization that would go beyond the engagement of Intel as a contributing member.

Intel's push into the WiMAX market encouraged other chip vendors such as Fujitsu, TeleCIS, Sequans and picoChip to introduce products as well. While it was convenient for many vendors in the WiMAX Forum to initially give Intel the leader's role, they are now starting to step up and lead, contributing to the marketing and customer/channel education. Now that there are certified products actively competing for service provider attention (see also Bell South's decision to trial Alcatel's mobile WiMAX products), vendors will drive the messaging and marketing forward. The WiMAX Forum would be well-advised to monitor the situation, emphasizing the significance of interoperability and certification to keep the momentum of WiMAX going.

Thanks to Intel and its huge effort to get the broadband wireless market off the ground.

Submit a Comment | Permalink

The big buzz in Denver last week at the Cable-Tec Expo was all about cable bandwidth - how to get more of it and how to use what you have more efficiently. Cable TV systems are increasingly being tasked to support expanded programming content for video services and higher speeds for broadband services. Primary among bandwidth-hogging services is local, national and on-demand high-definition (HD) content. HD changes everything in Gartner's view, and in order for cable to match or exceed the HD offerings by satellite operators and telco IPTV contenders, cable's got some tinkering to do in their networks. There was less talk about how to support ultra-high-speed Internet access up to and beyond 100 Mbps and business services, which requires a move to the next-generation standard for broadband - DOCSIS 3.0, as well as implementing bandwidth overlay technologies that can support commercial services in spectrum up to 3GHz versus the usual 750/860MHz of spectrum most cable systems operate within.

The key technologies that were being showcased in Denver included switched digital video technologies from RGB Networks, Big Band Networks, Terayon, Scientific-Atlanta (SA), and others; HFC expansion; and overlay technologies from Vyyo, SA, C-Cor, and Narad Networks that grow cable's spectrum beyond 750/860MHz up to and beyond 1GHz.

On the high-speed data front, Arris interestingly introduced a proprietary solution expanding cable data modem bandwidth to 100 Mbps over the HFC platform by bonding multiple channels together. But the Arris solution offers no guarantee that its "FlexPath" technology would be usable when the DOCSIS 3.0 standard is completed, and it is really targeting the solution to cable operators (mostly in Asia/Pacific and Europe), which are in fairly desperate straits when it comes to competing with providers offering speed from either VDSL+ or FTTx-based networks.

Note on vendor consolidation: In light of the larger telecom mergers happening around the world, it wouldn't be a cable show without a modicum of buzz on the floor on who was buying whom. Motorola was frequently being paired with VOD technology supplier Broadbus, Cisco Systems with another VOD vendor called Arroyo, and finally and more definitively, Arris with Concurrent Computer. Gartner believes that the Arris pairing with VOD company Concurrent is the most likely. Arris is strong in telephony and high-speed data, and it is the top supplier of VoIP terminal adapters with embedded cable modems; it has been ranked either second or third in the global cable modem termination system (CMTS) business. It's been pushing an edge QAM solution and a next-generation CMTS that enables more efficient processing of streaming video over the HFC platform, but the company's core DNA is not in the video business per se, so any acquisition that strengthens its ability to bring a more holistic video solution to telco and cable operators would help position it better.

Terayon - which has a strong video processing business - is still struggling with its future, and sources told Gartner that most potential buyers have walked away from considering any acquisition until Terayon cleans up its books and finally reports its financials. Many believe the asking price is too hefty, considering the earnings for its main video product line - the CherryPicker. Portions of C-Cor were mentioned as being on the block, which includes a VOD unit. In addition to the VOD vendors, other contenders mentioned as ripe for acquisition were the encoder vendors: Harmonic, Tandberg Television and Modulus Video, among others.

Submit a Comment | Permalink

Here and here, we have we have two of the most recent entries on the latest Silicon Valley game-changing company, YouTube. Since YouTube was launched less than a year ago, it has quickly become the focal point on the 'net for those who create video content and want to publish it or those who find it online and want a way to share it.

Implications for network providers could be significant. When millions of people start uploading content to YouTube - content they've either created themselves or patched together using (probably) copyrighted material - there is bound to be some effect on ISPs and their network traffic. Of course, how soon we have millions of people uploading content is anybody's guess. On the other hand, YouTube is averaging 50 million page views a day and as many as 50,000 video uploads per day.

While YouTube is certainly the most visible, a number of other sites are also generating huge amounts of video-related traffic. One, a social-networking site that also carried video downloads of a major surf contest held last winter, generated 8 terabytes of video traffic in one 24-hour period.

Conventional landline networks won't be alone in dealing with the onslaught of video: mobile phones with (still) cameras are de rigueur and video-capture capabilities are starting to crop up. Consumer appetite for online (and eventually wireless) video appears endless; can the carriers keep up? (see "Super Enablers Begin to Reshape the Media Ecosystem.")

2 Comments | Submit a Comment | Permalink