In late 2005, we published research (see "Gigabit Power-Over-Ethernet Switches Are Largely Unnecessary")stating that deploying gigabit to the desktop would result in $10 billion of wasted expenditures during 2006 to 2008. Our research expected that the majority of this waste would occur in 2006, that prices for workgroup switches supporting Gigabit Ethernet would decline, and that the delta between 10/100 Ethernet and Gigabit Ethernet would narrow substantially. We believed that the opportunity for savings would be significantly diminished beyond 2008.

Looking back at the last 18 months, we see two substantial things have happened. First, most high-volume gigabit products have not declined in price. Second, we have seen a resurgence in the 10/100 market with new enterprise-class switch families from Extreme Networks and Nortel Networks. At this point, we believe that organizations that procure workgroup gigabit switches will continue to waste significant resources for the foreseeable future.

Looking through our recent quarterly statistics for enterprise-targeted Ethernet products, we find that the delta between workgroup gigabit switches and 10/100 is $53/port. Put in other terms, an average copper gigabit port costs more than 2.5 times the price of an average 10/100 port. The waste in the past quarter alone was $800 million (or more than $3 billion on an annual basis). With the price delta remaining as high as it is, we're seeing that our $10 billion prediction underestimated the waste by a significant amount.

We're in the process of updating this research - not only in the Ethernet market, but in other network and communications requirements as well. Our new research will be presented at our Enterprise Networking Summit in early December in Nashville.

Submit a Comment | Permalink

Sony, in conjunction with videoconferencing distributors Imago and Teleportal, recently announced the launch of Sony 3D Telepresence for the U.K. market. The Sony Telepresence solution is described as utilizing three-dimensional (3-D) technology to display participants as life-size in appearance, live, within an apparent 3-D environment. The Sony HD Video Communication (PCS-HG90) system's built-in MCU provides true high-definition (HD), two-way communication in real time between up to four sites. The PCS-HG90 supports H.264 HD over standard IP networks up to 8 Mbps to deliver high-quality, HD video 1280 x 720 resolution at 30P and 60P. The system also offers wideband stereo sound. The price tag on a single, HD roll-about system with a 50-inch screen (available now) is about US$60,000. Sony Telepresence room configurations, specifically designed for the high end of the telepresence market, are expected to become available in fall 2007. No word yet on availability beyond the U.K. market.

Sony is a late-comer to the telepresence market, joining vendors such as Cisco, HP, Polycom, Tandberg, Telanetix, Teliris and others, but it could make a significant impact based on its high-visibility brand name and great reputation for technology. If it can also deliver the Sony Telepresence solution at a lower price point than the competition - an issue for many video customers who feel that the cost of telepresence in general is an inhibitor to adoption - that could be a key differentiator for the company's success in the telepresence market (see "The Gartner View on Enterprise Video", "Toolkit Tutorial: Buying Into Video Telepresence", and "Hype Cycle for Enterprise Communication Applications, 2007".

Submit a Comment | Permalink

HP and Tandberg recently unveiled a gateway solution that will link HP's telepresence system (Halo Collaboration Studios) with standards-based videoconferencing systems from Tandberg. The HP Halo Gateway will help organizations meet requirements for bridging some existing Tandberg room-based videoconferencing systems with HP telepresence environments. The solution consists of a Tandberg 6000 MXP codec, a Tandberg Video Switch, a Halo compositor, an HP ProLiant server and HP-designed software. It stems from an alliance Tandberg and HP announced in January to make their respective videoconferencing and telepresence portfolios interoperable. The alliance's goal is to enable customers to collaborate using a range of products - from the immersive telepresence experience of Halo Collaboration Studios to Tandberg's conference room and desktop-based systems - without requiring further investments in internal IT resources or upgrades to corporate networks.

Of the organizations that Gartner has talked to about telepresence, many have indicated that the ability to leverage existing investment in their videoconferencing systems at off-sites by interoperating with telepresence systems at main sites was a requirement. The ability to interoperate is starting to be addressed by several of the telepresence solutions on the market, with some supporting it already. Overall, interoperability is certainly helping to round out the functionality of the still-maturing telepresence market.

For more information, please see: "Toolkit Best Practices: Building the Business Case for Telepresence," "The Gartner View on Enterprise Video" and "Hype Cycle for Enterprise Communications Applications."

Submit a Comment | Permalink

In what has been a long drawn-out process, it looks like Motorola has emerged the winner in the contest for Santa Clara-based Terayon. Motorola reported that it will pay $140 million, or $1.80/share for Terayon, whose business is in digital video processing, including digital ad insertion systems and CherryPicker video processing system.

Gartner believes Motorola won for a couple of reasons, but mainly because two of its top executives - CEO Jerry Chase and Matt Aden, senior vice president of global sales and customer support - were both ex-General Instrument/Motorola alumni. Better the devil you know than the devil you don't, I say. It was also likely that Motorola's stronger financial resources beat out the likes of Arris and Harmonic. Some equity analysts are already downgrading Arris as a result of losing this bid. Motorola had big incentives to keep Terayon out of now arch rival Cisco/Scientific-Atlanta's talons.

In addition, while Motorola has been partnering with RGB Networks for its highly regarded and powerfully elegant switched digital video (SDV) solution that many cable operators are now installing to make room for massive amounts on-demand and HD content on their networks, Terayon has its own version of a switched digital solution. It remains to be seen what will become of any future contracts for RGB's SDV gear once the deal is closed. In any event, many financial analysts believe, and I concur, that RGB would be a more expensive acquisition than Terayon.

Terayon joins a host of other recent small yet tactical acquirees Motorola has picked up over the past year or so, building a broad portfolio of "connected home" technology data and video solutions over IP, RF and ATM-based networks: Kreatel for IPTV set-tops and software; Netopia for DSL CPE, Broadbus for VOD systems; Tut Systems for IP-based video encoding and processing; Vertasent, which specializes in software for switched digital video network technology; NextNet Wireless for WiMAX/broadband wireless access systems; and Ucentric Systems for multiroom media networking software.

Submit a Comment | Permalink

The recent VoiceCon held in San Francisco highlights the commoditization of IP PBX handsets. The user community is resisting high-end, proprietary handsets with large LCD screens. Vendors in turn are slowly acknowledging that their plans for profitable $300-$400 handsets will not manifest.

As enterprise voice over IP (VoIP) is implemented, users are instead relying on their desktop GUI because of: 1) their experience with the desktop; 2) the desktop's large display (much larger than that of a phone LCD); 3) Microsoft Office integration; and 4) reduced costs. These factors are yielding a "dumbing down" of the IP PBX handset (see "Don't Purchase IP Screen Phones If You Have a PC on Your Desk").

Cisco, Avaya and Nortel are grudgingly moving to more generic, standards-based, SIP phones. Independent handset vendors like Polycom and LG offer basic SIP phones in the $150 range. Future price pressures are inevitable as the SIP standard matures. The unfortunate byproduct of this environment is that the current base of IP phones is as "stylistically challenged" as the 1970s relics they are intended to replace.

1 Comment | Submit a Comment | Permalink

Here's a good example of how properly designed and managed wireless can be more intelligent, and more secure, than wired networking.

Wireless LANs are becoming a standard part of most enterprise networks, expanding to cover entire facilities - more than just conference rooms. Next, we'll see applications that take advantage of that infrastructure. As an example, device location tracking is just now becoming an affordable and integrated feature of WLAN infrastructure. This information has a wide variety of uses: It could allow you to restrict usage of applications to certain areas in a building. It could alert you when devices leave a certain area. You could create a graphical map with the physical location of every Wi-Fi device in your facility. Applications could prove useful in manufacturing, logistics, healthcare and many other industries.

Newbury Networks recently announced a software development kit (SDK) for developing location-driven wireless LAN applications. We're still in the early stages of this technology, and these applications will require thorough calibration of location signatures for tracking applications to work accurately. But this is a clear indication of where this technology can go. For more information on Newbury Networks, see here.

Submit a Comment | Permalink

While we still consider wide-area network optimization controllers (WOCs) to be an innovative technology, they can be too expensive for very small offices. Intelligent Compression Technologies provides similar functions using client software that it calls AcceleNet, which prices out at just a couple hundred dollars per user. But it promises to improve performance by over 10x!

See "Cool Vendors in Enterprise Communications, 2006".

Submit a Comment | Permalink

As networking technologies became more affordable over the past decade, the world moved from connected companies toward connected individuals. Twenty years ago, companies connected their own locations, and then they connected to each other over private value-added networks (VANs) and the Internet. Over the same period, the mobile phone moved from being an extravagance for the rich to become an essential tool (and a fashionable lifestyle accessory) for individuals in almost every economic band in almost every global market.

In the coming decade, we will move toward the connected world. Cheap, network-enabled, self-organizing sensors will emerge. Sensor nodes will cost less than $10 by 2008, and prices will plummet from there. Gartner predicts that sensors will enable the real-world Web, providing information to enhance business and personal decision making.

The implications of the real-world Web are huge: Today, a large-scale network may have (at best) millions of nodes; sensor networks will have billions. Connected objects will be everywhere, be on everything and generate millions of real-time data points. Such ubiquity will demand a shift in software and hardware to what Gartner calls the "tera-architecture" - processing terabytes of data every second. This will spur an event-driven architecture that enables sensor networks to respond directly to their environments. Multiple architectures will exist for sensory applications as well.

Communications network planners might actually have a reason to think about IPv6, which to date has been a nonstarter. But don't do anything now: Even the U.S. government is starting to waffle on its commitment to IPv6. For the vast majority of companies, IPv6 will not even be on the radar screen before the end of the decade. Unless you plan to lead the way in sensor adoption, there are better networking investments to make.

Submit a Comment | Permalink

Innovation can be about technology, or it can be about what you do with the technology. There are also innovative ways of organizing your operations to anticipate change in your company.

To take an example, voice services are certain to become integrated into data applications. Telephony not only will be for making voice calls, but also will be tightly entwined with collaborative applications. So how should voice organizations adapt? As noted in "How to Structure a World-Class Organization", we believe that, in the next couple of years, voice people (especially those that focus on endpoint issues and user support) will move out of the networking organization and into application and desktop teams.

Security support is following a similar trajectory. Gartner believes that best-practice security teams will be moving outside of IT, leaving network and server support groups to manage security operations, with oversight from the security team.

Submit a Comment | Permalink

The classical approach to controlling malicious code (meaning viruses, worms and other malware) uses databases of code signatures. But a single attack can easily mutate into hundreds or thousands of variants in a very short time - and each mutation requires its own signature. At this rate, signature databases will collapse under their own weight. Emerging tools cluster all variants into a single class, examining the underlying behavior instead of the signature, and as a result, they handle the mutation problem much more efficiently.

Prevx takes this approach and then applies "networked collective intelligence" to supply and gather data in real time over the network. It uses the power of growing network value (also known as Metcalfe's Law), which means that every new user added into the system improves the experience of all existing users. It's similar to Cloudmark's successful approach to spam and virus control in the consumer and ISP markets. We believe that Prevx is in a good position to become a partner of Cisco, ISS or Webroot - potential even an acquisition for them. See "Cool Vendors in Security and Privacy, 2006."

Submit a Comment | Permalink

Supporting remote offices has never been easy, but it's become much harder in recent years. In the traditional model of IT delivery for remote offices, applications sit on local servers in the branch. More applications mean bigger servers and more software licenses. As the number of remote locations increases, more locations mean more servers. Each server requires a lot of care and feeding - and lots of money, in the form of capital expense, software licensing, maintenance and support. This approach we call the "persistent branch," a model that is optimized for high performance and availability.

But there is a new approach, which we call the "persistent data center", in which services are centralized over the cloud. It optimizes for cost and control, but it doesn't mean you have to give up user performance - at least if you deploy a new class of innovative devices, which we call "wide-area optimization controllers". These devices understand critical applications and optimize popular network and file system protocols, while providing features like quality-of-service prioritization, compression and caching.

But to deploy the full-service branch, you can't just think about user applications. There are a number of infrastructure services that you need as well. There are logical elements like policy, route control, DNS and directory issues, which are essential concerns for the branch too.

As a result of all this, we believe that a new class of devices will emerge, something we call, simply enough, "branch-office boxes". These boxes not only serve as WAN optimization devices, but also provide DNS, DHCP, printing services and Active Directory caching.

While multifunction router devices such as Cisco's ISR, Juniper's SSG and Nortel's Secure Router have added integrated features, especially security, it's only a start. And it's not only the networking companies that will compete here: Others will introduce their own branch-office boxes. The persistent data center is a threat to the server and software vendors, who stand to lose a bundle. So it's not too hard to imagine IBM, HP and, yes, Microsoft playing their hands in this game too.

1 Comment | Submit a Comment | Permalink