Top Ten Vendors to Watch
Now it's time to look forward, with yet another Top Ten list, this time identifying companies worth watching in 2003. The term "worth watching" means just what it says, by the way. This is not Light Reading's Top Ten Private Companies – it's a list of companies that appear to have a lot of potential, but have yet to prove it. What we'll be watching for in 2003 is a breakout: product details, beta trials, customers announced, or, as is too often the case these days, a promise broken. The product ain't all that hot after all, and once the covers are taken off, we can all have a good snicker.
In order to capture the most interesting vendors for this list, I've cast my net wide – asking all of the editors of Light Reading, Byte and Switch and Unstrung to chip in with their choices. Here's the result:
10. Ellacoya Networks Inc. Ellacoya is a bit of a puzzle. They made our Top Ten Private Companies in December 2000, then fell off the list, seemingly into startup oblivion. But they’re back, and with a new, focused mission: not so much service creation as "service control" (see Is Ellacoya on the Comeback Trail?). It goes like this: ISPs have a love-hate relationship with peer-to-peer traffic. It makes people avid users of services and inspires them to upgrade to broadband speeds, but it swamps the network with traffic that is hard to manage and impossible to bill for. Recent data from Tier 1 U.S. MSOs (multiple system operators) shows that during the prime-time evening hours, 40 percent of their user base are using peer-to-peer (P2P) applications, consuming up to 70 percent of the total network bandwidth. This pains ISPs greatly, and right now they often just shut off a user if they are considered "abusive" – but that doesn’t build customer loyalty, nor does it really solve the problem. Ellacoya believes this is a problem that can be solved, and it has built a system that provides transparent control of network traffic by application (through proprietary application fingerprinting) at multi-gigabit speeds. This allows broadband operators to promote network "fairness" by controlling how much bandwidth is consumed by these robotic P2P applications, thus ensuring that "interactive" services such as Web and email are not squashed. Ellacoya says that its flow-based architecture enables operators to gain full visibility into ALL of their network traffic by application, user, and destination. Ellacoya’s switch supports passive user authentication, tying a user to a particular IP address even in a fully dynamic DHCP (dynamic host control protocol) environment. This service control is the Trojan Horse: Once inside the gates, Ellacoya will then be positioned to upsell operators on additional application suites such as security, billing, tiered service packages, and so on. Will ISPs let Ellacoya inside the gates? That’s what we’ll be watching this year. 9. Timetra Networks If Procket Networks Inc. was the router company to watch last year (not that much actually happened), then Timetra is worth our notice this year. They’ll finally be telling the world what they’re up to, and just how a "service router" differs from today’s edge routers (see Timetra Trumpets Its Edge Router ). It’s common wisdom in the startup world to say, "Hey, we’re not competing with Cisco or Juniper, we’re complementary!" But that’s really just to assuage VC worries over fighting Goliath with a sling. Timetra, as far as we can tell, is going after those guys – there’s really no avoiding them when you come out with an MPLS edge router. In 2003, as the market for IP services continues seek its shape and value, Timetra will add another box to the mix. Everything we hear so far is great. This box has everyone impressed, and it doesn't hurt that Basil Alwan is a gerat guy who does not seem affected with enhanced entrepreneurial ego syndrome. 8. BigBand Networks Inc. Cable MSOs are getting serious about multimedia. This may be the year (finally!) of video on demand – and not just movies on demand, but real interactive multimedia services, delivered right to your living room or PC, at prices that make sense and using equipment that really works (see IP Video Uptake by 2004?). Cable’s entry into multimedia has been painfully slow, and many a startup has had dirt shoveled over them for trying to help MSOs get into the interactive business. BigBand isn’t one of those. It has raised $57 million over the past couple years, landed Cox Communications Inc. (NYSE: COX), Rogers Cable Inc., Blue Ridge Communications, AT&T Broadband (now part of Comcast Corp. [Nasdaq: CMCSA, CMCSK]) and Time Warner Cable as marquee customers, and continuously improved its Broadband Multimedia-Services Router (BMR) to support a wide range of applications – from ad insertion to digital video transport, grooming, and video on demand. It’s hard to believe that the complacent MSOs of yesteryear are getting serious about "on-demand" networking, but they are. In contrast to the hapless CLEC trying to undercut an ILEC on T1 prices, MSOs have something real to sell, and consumers show an insatiable appetite for all things video. 7. MeshNetworks Inc. Mesh Networks has been tipped for possible stardom by Unstrung editors, partly because of its potential in our fearless leader's Homeland Defense initiatives (see Unstrung's Top 25 Startups). Mesh Networks has developed a "radio-agnostic," wireless, peer-to-peer networking technology that can support wireless technologies such as Bluetooth and 802.11, along with the company's own quad-division multiple access (QDMA) radio protocol. QDMA supports VOIP, data, and multimedia wireless transfers. Using military technology, Mesh Networks has developed a networking system that doesn’t use basestations – instead, each device on the network becomes part of the relay system. The basic idea behind mesh networking is a system that uses a series of wireless gateways, or nodes, that pass data and share bandwidth among them, rather than sending and receiving data from a central point. The mesh approach is attractive in a number of ways, one of them being that the available bandwidth increases with the number of nodes in the network (see Mesh Gathers Momentum and Commentary: Packet Relay Radio to the Rescue?). A downside might be that it's tough to predict how many hops traffic will have to take before it reaches fiber or its final destination. The variations in network delays that could result might create quality-of-service problems with some types of traffic, such as voice and video. Formed in January 2000, the firm has "over $27 million" in funding provided by VCs such as Apax Partners and Redwood Venture Partners LLC. 6. Topspin Communications Inc. This year will undoubtedly be a better one for the data center than for the telecom operator. While the world of telecom continues to suffer from financial uncertainty, regulatory uncertainty, and antique business models in transition, the world of the large enterprise remains overwhelmingly in need of scale, efficiency, and flexibility. Achieving those goals has been accomplished to date through a process of virtualization – first of storage resources through the implementation of SANs, and now of computing, thanks to cluster computing and grid networking. As this process continues, applications will no longer be associated with a single processor or a file server; instead they will run on virtual systems, stitched together out of distributed resources within a data center or among many. In 2003, Topspin will be betting that the best way to knit processing and storage together is through a purpose-built switch with wicked fast I/O, running a new kind of control plane that understands how servers and storage systems need to communicate in order to support powerful virtualization of the data center (see Topspin Takes Off). Today, Topspin is gambling that InfiniBand is the best switch fabric for the job; others think Ethernet will carry the day. Topspin is hedging their bets by putting Fibre Channel and Ethernet ports on the switch and adding a layer of customized software to tie everything together. We’ll watch Topspin in 2003 to see if its gamble pays off, and whether "switched computing" becomes a buzz word. 5. Xponent Photonics. Little information about Xponent is available on the Web -- the startup doesn't even have a Website at present. But that could all be about to change... Reportedly, the company was founded by the same group of executives that previously founded Ortel, a maker of lasers and component for cable TV that got sold to Lucent/Agere. In fact, Xponent was founded the same month that Ortel got sold to Agere. Some say that the reason the startup is being so cagey now is because its executives signed a three-year non-compete agreement with Lucent. This agreement would expire in February 2003 -- coincidentally the same month that Xponent told Light Reading that it plans to make its debut. With the management team's experience of building a successful company, we feel sure that when Xponent does come out of stealth, it will do so with a bang. Read more: Xposing Xponent Photonics 4. Avaki Corp. Though not necessarily a telecom player, Avaki warrants our eye in 2003 because they are providing the necessary "glue" for grid networks – networks comprised of computing and storage resources knitted together to form quite powerful virtual computing or data processing systems. And you know how keen I am on grid networks, don't you? (See Watch for the Grid). Avaki, founded in early 2001 under the name Applied MetaComputing, received initial funding from Polaris Venture Partners and changed its name to Avaki in mid 2001. Dr. Andrew Grimshaw, founder and CTO of Avaki, initiated the Legion Project, a government-funded grid computing project, at the University of Virginia in the early 1990's. Avaki was formed to commercialize the technology derived from Legion. At a high level, Avaki’s vision is one of a ubiquitous, pervasive computing infrastructure, one that will enable secure and transparent access to all computing resources. To that end, Avaki delivers grid software solutions, also known as middleware, that bring together data, compute, and application resources from multiple locations and administrative domains for industries such as life sciences, manufacturing, and oil and gas. Although grid computing concepts, technologies, and deployments have been floating around for several years, 2003 will be the key year in the development of middleware software solutions. Grid middleware, analogous to protocol and management software in communication networks, is the key ingredient for many of the major grid initiatives announced to date. Avaki is an early technology leader in this nascent, growing market, and has formed strategic technology partnerships with the industry’s most influential grid computing players, including Hewlett-Packard Co. (NYSE: HPQ), IBM Corp. (NYSE: IBM), and Sun Microsystems Inc. (Nasdaq: SUNW). 3. IPWireless Inc. IPWireless is a wireless startup vendor that could upset a few apple carts in 2003. Its basestations provide wireless data access using standards-based technology based on UMTS, a 3G standard, but it is not the same as the kit being offered by the likes of LM Ericsson (Nasdaq: ERICY), Nokia Corp. (NYSE: NOK), or NEC Corp. (Nasdaq: NIPNY). Conventional UMTS systems use frequency-division duplex (FDD) or "paired" spectrum, whereas the IPWireless system operates over time-division duplex (TDD) or "unpaired" spectrum. Data-only TDD systems are more spectrally efficient than FDD systems, because they use just one channel for both upstream and downstream traffic, instead of two. Most operators with 3G licenses have the right to use unpaired as well as paired spectrum. IPWireless claims its kit can offer data rates of 4.5 Mbit/s – well beyond what would be possible with mainstream wide-area network technologies – and can be deployed for about a sixth of the cost (see IPWireless Does Mobile 3-Step). It claims it will reveal the names of Tier 1 carrier customers, and OEM deals with two of the top five wireless infrastructure vendors, in the first quarter of 2003. It already has customers, so its products are already being deployed commercially, and it recently raised a further $27 million. 2. Huawei Technologies Co. Ltd. Huawei certainly has to be watched in 2003. It has emerged as a powerful networking equipment supplier in China, causing worries for Alcatel SA (NYSE: ALA; Paris: CGEP:PA), Cisco Systems Inc. (Nasdaq: CSCO), Lucent Technologies Inc. (NYSE: LU), Nortel Networks Corp. (NYSE/Toronto: NT), and Siemens AG (NYSE: SI; Frankfurt: SIE), all of which have devoted significant resources to building relationships and selling gear into the large, rapidly growing carriers in China. And now – watch out – Huawei is entering the U.S. market, via its Futurewei subsidiary (see Has Huawei Got Cisco's Number? ). The typical response to the threat is "the products are one-offs and won’t stand up to ILEC testing" – which, to our ears, sounds a bit like famous last words. Huawei is in a position to make telecom systems and data systems at a significantly lower price than its North American or European counterparts, and to quickly offer those systems at rock-bottom prices in Western markets. They may not make it into the incumbents right away, but neither have startups, and there are plenty of other customers around willing to try them out. Huawei is also in the low-end router business, and will be targeting enterprise customers with products that are as little as half the price of existing Cisco gear. 2003 should be an important year for Huawei. They have an opportunity to win significant contracts within China, providing a powerful revenue source from which to address emerging markets in India, Russia, and parts of Europe, even while fighting Cisco and the incumbent telecom vendors here for what contracts exist. In light of all this, they are already becoming a coveted customer for North American chip vendors, such as Agere Systems (NYSE: AGR/A), and are posing a true threat to suppliers around the world. 1. Infinera Inc. 2003 will be a year to watch Infinera, which could be the next big thing in optical communications, potentially rendering the optical amplifier obsolete for DWDM architectures. Alternatively, it could be yet another optical chip company that combines a few interesting active and passive functions together, but fails to solve the "Holy Grail" issue of complete signal regeneration. Infinera has the all-star team, backers, and buzz right now to warrant the top spot on this list, as it’s going after the heart of the optical network with something potentially paradigm-shifting (see Infinera Shoots for the Moon). We don’t know if they’ll pull it off, but they have to be watched in the year to come. — Scott Clavenna, Director of Research, Light Reading