Mostly Fearless Forecasts

Picking which telecom equipment startups have the best shot at acquisition or IPO over the next 12 to 24 months requires facing a number of hard questions. Who will win the telco vs. cable battle? Can anyone really succeed with yet another core router or 10-Gbit/s Ethernet switch? Are all IP appliances doomed to become blades and feature sets on Cisco routers? Is this FTTP stuff for real? Will WLAN switch vendors make any money from telcos? After a while, the head spins; each one of these questions is a report in itself.

But analysts love matrices, taxonomies, and, most of all, lists. So in Telecom Recovery Investment Opportunities, we applied a bit of science and a lot of subjectivity to picking the top startups for 2004. First we broke the telecom equipment market into 21 categories, grouped public and private vendors into each, and then evaluated the prospects for growth and differentiation for each of those categories. The ones we ended up liking included the obvious Ethernet access, the nebulous high-value IP, the risky cable/MSO advanced services equipment, the slightly out-of-bounds WLAN switch market, and the confusing broadband remote access server (B-RAS) sector.

If there is a thread holding these all together, it is an acknowledgement that access is king in 2004: Operators want to use access equipment as a way to finally add some value to their IP service portfolios.

We created a scoring system for these categories, awarding each segment one to five points for its ability to demonstrate rapid market growth, the presence of weaknesses in incumbent product lines, the quantity of VC investment to date, the opportunity for a new startup to differentiate itself, a low cost to capitalize, and at least some indications of a path to acquisition or IPO. The categories listed below fit the bill.

Table 1: Top Market Segments for M&A and Investment
RANK Market segment Rapid market growth? Incum-bent weak-nesses? Few start-ups? Oppt'y to differen-tiate? Low cost to capital-ize? Clear, high-value exit? SUM
1 B-RASs 4 3 5 5 3 4 24
2 Cable/MSO Advanced Services Equipment 5 4 3 5 3 3 23
3 IP Service Controllers 5 4 2 5 4 3 23
4 Ethernet Access Equipment 4 3 3 4 4 4 22
5 (tie) Carrier Packet Voice Infrastructure 5 2 2 4 3 3 19
5 (tie) Multiservice Switching (FR/ATM/ MPLS) 3 3 5 3 2 3 19
5 (tie) 802.11 Wireless LAN Switches 4 4 3 3 3 2 19
Source: Heavy Reading

When it came to picking private companies to single out, a great deal more subjectivity was applied. Many of our picks had in common a record of revenues, a projection for a real "up" year in 2004, partnerships with leading OEMs, and a place in relatively hot market segments. Finally, they demonstrated that major incumbents were not entirely bulletproof.

But along with the obvious selections, we have a few odd choices that are worth explaining in more detail. First there's BigBand Networks Inc., a supplier of an edge router designed to deliver advanced video services, including video on demand, switched digital video broadcast, and rate-shaped HDTV delivery. BigBand's success as a startup in the MSO space is impressive.

VCs have historically shied away from this area. From our interviews with VCs for this report, the note sounded on cable was strikingly uniform, regardless of whom we talked with. Cable companies are notoriously cheap and ruin margins for private companies. The supplier oligopoly of Motorola Inc. (NYSE: MOT), Scientific-Atlanta Inc. (NYSE: SFA), and Arris Group Inc. (Nasdaq: ARRS) make it a real pain to break into the top tier of operators. CableLabs is a negative force in the industry, devising specifications with the singular goal of commoditizing equipment. Add the fact that most VCs don't have the kind of long-cultivated contacts in this segment that they have with North American telcos, and you have a situation in which cable is considered a rather nasty foreign country to invest in, rather than a fertile, innovative market.

But BigBand is succeeding. It can count Time Warner Inc. (NYSE: TWX) and Comcast Corp. (Nasdaq: CMCSA, CMCSK) among its customers, which is absolutely a must if you're going to prevail in this strange land. It is attacking an application space in which traditional suppliers are notoriously weak, namely VOD, IP video, and digital broadcasting services. All of these services require a significant level of media processing and switching, something big suppliers are not used to.

Beyond the cable market, BigBand has set its sights on telcos as they make their own transition into video. This is particularly attractive ground for BigBand, since telcos aim to differentiate from MSOs through more interactive digital services. Telcos understand switched services much better than MSOs do, and must push forward with these from day one to compete. Finally, we like BigBand's VCs, particularly Santo Politi of Charles River Ventures, who knows the cable space well and knows how to get companies into the top tier.

Next, Luminous Networks Inc. needs some explaining. Luminous made the list, easily fulfilling most of the requirements for success (OEMs, diverse customer base, good market segment, increasing revenue), but it certainly has been a tough company to love over the years. For starters, you have to write the name of its marketing VP in pencil. You also have to believe that RPR (resilient packet ring) is for real, that growth in China and Latin America will translate into success in North America, that an OEM relationship with Ciena Corp. (Nasdaq: CIEN) is a positive, and that the level of competition in the MSPP space will allow a private RPR vendor to make it.

The good news is that Luminous has raised plenty of money to make a solid showing in the metro market worldwide, and it was savvy enough to follow the activity in the cable market first, and the telco market second. The question hanging over Luminous is just how it will exit and return some value to investors. An acquisition by Ciena is the obvious choice, but the Scientific-Atlanta OEM deal complicates the picture. IPO is a huge longshot for Luminous, given that Wall Street will need some real convincing to support an MSPP supplier in a market so utterly dominated by the likes of Alcatel SA (NYSE: ALA; Paris: CGEP:PA), Cisco Systems Inc. (Nasdaq: CSCO), Fujitsu Ltd. (OTC: FJTSY; Tokyo: 6702), Lucent Technologies Inc. (NYSE: LU), and Nortel Networks Corp. (NYSE/Toronto: NT).

Just saying that list out loud sort of deflates one's enthusiasm rather quickly. Luminous made the cut, though, and we're sticking with it.

Then there's Force10 Networks Inc. Force10 appeared so perfectly situated in 2002 that its success seemed all but inevitable. Every customer that got its hands on a Force10 switch loved it. It delivered what it preached: line-rate performance at 10 Gbit/s, with all the features turned on. Add to that its status as first to market with a 10-GigE switch, and you'd think an incumbent vendor would have taken Force10 out just for lack of any other quality startup out there and a near-guaranteed market opening as service providers and enterprises migrated GigE infrastructure to 10-GigE.

But sometimes the world doesn't cooperate with pat predictions. The market slows down (or grinds to a near halt, depending on where you look), and the big incumbent vendors scramble to catch up. And catch up they did.

Cisco got 10 GigE out the door and slashed prices, as any good market dominator should. Then Foundry Networks Inc. (Nasdaq: FDRY) followed suit, delivering an edge and core 10-GigE switch and pushed prices down even further. One day you're first, the next you're eating sand off the heels of your rivals, which not so long ago you had said were caught sleeping.

Force10 may no longer be in the pole position, alas, but we still like its prospects. Its boxes remain rock solid, no matter whom you ask, and it keeps winning customers. The big question is not whether Force10 will find an exit (this is the Ethernet switch market, remember), but whether the investors will make out. To date, Force10 has raised over $200 million, rivaling core router startups and other money pits for the title of "most VC capital still at work today."

That said, we're more inclined to believe in an IPO for Force10 than an acquisition. Why not? The Street can see the value of a company that straddles the enterprise and service provider worlds well, and Force10's only real weakness is the fiercely competitive world it inhabits, with its pricing and margin pressures. But Force10 has proven it can compete, so why not set it on the waters of the Nasdaq and see if it floats?

— Scott Clavenna, Chief Analyst, Heavy Reading

On April 13, analysts from both the Light Reading Insider financial research service, and Light Reading's Heavy Reading market research division will examine technology investment questions like those addressed above at a unique event hosted at the W Hotel New York Union Square in Manhattan.

The conference –
The Telecom Recovery: Opportunity Amid the Chaos – will examine the investment potential of a wide variety of technologies including triple play, 802.11 WiFi, FTTP, hosted SIP applications, SSL VPNs, IP video, MPLS-based convergence, and carrier-class Ethernet. It also will feature keynote speeches by Hossein Eslambolchi, President of AT&T Global Network Services, Pradeep Sindhu, CTO of Juniper Networks Inc., and Steven D. Levy, Managing Director of Lehman Brothers.

Register here.

IP Observer 12/5/2012 | 2:12:56 AM
re: Mostly Fearless Forecasts There will be no specialized BRAS devices by end of year. Functionality will be integrated into existing devices. Vendors are trying to create a specialized market in order to differentiate themselves from market leaders.

This is akin to what happened with Cosine when it tried to create a separate market for its products. What happened? Well, category disappeared.
Scott Clavenna 12/5/2012 | 2:12:56 AM
re: Mostly Fearless Forecasts That's a great point. Functionally, the BRAS has risen in importance, but exactly where that functionality resides remains debatable (edge router, DSLAM/CMTS, or standalone). It appears integrating BRAS into the DSLAM is the least desirable choice, with the remaining two options having their merits depending on the operator.

The odd thing with this ranking is that BRAS came out on top, but there really weren't any startups to point to as leasders in this space(Laurel, sort of, with its BRAS functionality, but it remains an edge router first and foremost) nor was there evidence of many new entrants on the way.

This functional segment, therefore, has the merits of a top category for further investment according to the scoring system, but product evolution may be obsoleting the opportunity by subsuming the BRAS into other systems.

jamesbond 12/5/2012 | 2:12:51 AM
re: Mostly Fearless Forecasts Can somebody describe what does BRAS do in
technical terms and not marketing lingo.

Scott Clavenna 12/5/2012 | 2:12:50 AM
re: Mostly Fearless Forecasts One place to look is the DSL Forum website, and read their technical reports TR-058 and TR-059, which describe in detail the evolving role of the BRAS in DSL networks.

bigpicture 12/5/2012 | 2:12:47 AM
re: Mostly Fearless Forecasts The BRAS came from concept of a Broadband RAS. As such it is essentially the managed entry point for broadband subscribers into the network. Back when Redback essentially defined this market there were short comings on the other network elements where this platform was located. The edge ATM switches aggregating the DSLAMS couldn't support enough VCs for all of the subscribers and the edge routers couldn't support enough logical interfaces or had the processing horsepower to terminate all of the PPP sessions. The BRAS could aggregate all of the ATM VCs and terminate a large number of PPP sessions or logical IP interfaces(>10K).

The BRAS supports a number of functions depending on the chosen deployment model. For PPP based deployments like PPPoE or PPPoA it serves as the PPP termination point, first logical IP hop into the providers network, supports subscriber login via RADIUS and service definition point for things like L2TP tunneling, NAT, Web portal redirection, some like Cosine and Shasta try to do VPN tunneling, firewalls and QoS at that point.

Most BRAS also have some ability to apply a "session" type of management for non-PPP deployments over 1483 B/R or Ethernet.

The key is that the BRAS spot is essentially the first layer 3 hop in the providers network. It is the logical place to apply IP services. TR-058 and TR-059 define many of those services.

Routers have come a long way since 1998/9. Juniper's ERX and Cisco's 10K can now scale in number of IP interfaces and PPP sessions as well as support higher densities of ATM VCs. Providers no longer need a specialized BRAS just to connect and aggregate their broadband subscribers. The BRAS platforms with firewall and VPN support will still find some markets for those who want to sell advanced IP services, but they aren't needed to solve the access problem any longer.

The DSLAM vs. Router debate revolves around how many edge IP elements do you want to manage. There are a lot more DSLAMs than edge routers. My observation is that carriers prefer centralized architectures.

jrturner 12/5/2012 | 2:12:44 AM
re: Mostly Fearless Forecasts
Regarding BRAS functionality and whether or not this functionality should be integrated into access equipment (like DSLAMs), I believe the ILECs would like to see as few provisionable elements in the network as possible. Further, the DSL Forum is closing ranks (driven primarily by ILEC requirements) around a new set of working texts (the WT-8x series), that recommends dynamic, application-level control over traditionally "hard-provisioned" network attributes, specifically quality of service, and even more specifically, IP quality of service.

In order to achieve this level of functionality, end-to-end, the network access providers (NAPs) and the ISP or ASP, are going to have to be on the same page with regards to network element capabilities, including dynamic application-level protocols signalling protocols that turn on/off individual element traffic engineering capabilities per application stream, at the time the stream is turned up.

If any network element between the stream source and the stream sink is not aware of these protocols or IP routing capabilities, then the service is subject to failure.

The current direction is to make the DSLAM much more intelligent with regards to these capabilities, essentially borrowing most, if not all, of the IP QoS/routing/tunneling capabilities currently found in Redback, or other BRAS-type of equipment. There are less elements to manage and provision, and it prevents classical ATM-only DSLAMs from aggregating packets onto ATM virtual circuits without regarding to IP quality of service.

There's alot more to this topic than could be included in a message board, but it's important to mention that these next generation technical recommendations coming out of the DSL Forum (whether they realize it or not) will also map nicely into FTTP networks now either being trialed or planned by Verizon, SBC, and Bellsouth.
Sign In