This was a year of refreshing pragmatism, so we’re looking more at important progress here, rather than newness. That said, this was also a year of coming of age. Many of the selections on our list have had a rough adolescence, starting out rosy and full of enthusiastic promise - only to be slapped down, laughed off the playground, and not allowed to join in any reindeer games. But now they’re back, well-muscled and mature.
Here's our picks:
10. Network Processors Come of Age
We had to find something positive to say about something in the components market in 2002, and this is the best we could do.
System vendors are beginning to take network processors seriously at last - partly because they actually exist now (making a nice change from the hype of a couple of years ago) and partly because OEMs are under pressure themselves to buy in components rather than developing their own.
Having said that, the future is not exactly rosy for companies in this space, because there's way too many of them - about 30, according to Linley Gwennap, president of The Linley Group, a consultancy specializing in network processors. A shakeout appears inevitable.
Notable articles on network processors this year:
9. Security Becomes a Hot Topic
Between hackers and terrorists, it’s a miracle anyone turns on their computer in the morning, but we all do - typing in credit card info, logging onto corporate VPNs, and running Websites. Networks are insecure and equipment is vulnerable, so security must continually evolve with the ways in which networks are used.
The ongoing process of migrating many network services from traditional circuit or Layer 2 networks to IP networks requires new forms of security, keeping this topic a hot one year after year. 2002 was no exception.
This year the government got involved, putting money into network security to shore up its own defenses. Startups began repositioning around security to take advantage of the interest, and NetScreen pulled off an IPO when no one else could.
Many large enterprises started distributing data center resources to new locations, so no one event could bring down the whole network, while carriers rolled out “managed security” services for customers unable or unwilling to trust their own.
Notable articles on security this year:
8: Video on Demand Comes Back into Vogue
This has been on top ten lists for a decade now, but this year we think it’s actually earned it. Video may not save the industry, but consumers appear ready pay for choices in their programming, and the technology around on-demand delivery of video has matured to the point that it doesn’t cost an armchair and a leg. Encoders have gotten cheaper and easier to use, Gigabit Ethernet has shown up as an interface on video servers, and now set-top boxes are available for under $300, making this service palatable to even the most parsimonious MSOs.
What’s behind VOD in 2002 that wasn’t there before:
- MSOs have significant digital subscribers in place. They added 5 million in 2002 alone, for a total of 19.5 million in the U.S. Having digital communications with the subscriber is tantamount to offering any advanced service. So, whereas in the past you had to bring a whole new infrastructure to a subscriber to offer VOD, today it’s merely an incremental step.
- Gear is cheaper, plain and simple, and it was the cost of infrastructure that made VOD services unprofitable in the past. Not so today.
- More variants. It’s not just movies on demand, but subscription VOD, which lets you watch past episodes of the Sopranos for a fee. Very popular, it turns out.
- Cable Television Laboratories Inc. (CableLabs), the technical specifications arm of the industry, adopted the first set of specs for standardized VOD infrastructure in March of this year. These specs are meant to tie set-top boxes, program-guide software, and billing systems with VOD equipment, hopefully creating a plug-and-play infrastructure for MSOs wanting to offer VOD.
Notable articles on VOD this year:
7: MPLS Gathers Momentum
Multiprotocol Label Switching (MPLS) had to be on this list because many of the big vendor sales centered on it, suggesting that the technology is establishing itself as a key part of carriers' core network infrastructure.
This was also the year that the term “Martini Draft Compliant” began to really mean something. Startups like Laurel Networks Inc. made hay by producing a fully Martini-compliant box, and others are following suit. The Martini Draft outlines how Ethernet and various Layer 2 protocols can be transported over an IP/MPLS core simply and easily, making the integration of Ethernet metro networks and carrier IP core networks a reality in 2002.
Fast Reroute also appeared this year, offering a bandwidth-protection scheme positioned as a cost-effective alternative to Sonet/SDH protection.
Notable articles on MPLS this year:
6: IP Reliability Becomes a Must-Have
This was the year when router vendors launched major efforts to improve the resilience of IP networks, with the goal of attaining the magic "five 9s" (99.999 percent) availability often cited by telecom operators.
These efforts go beyond having fault-tolerant hardware. They focus on avoiding data loss and minimizing delays when router networks reconfigure themselves around line or equipment failures, or suffer "flapping" caused by router processor problems. Terms such as "graceful restart" and "stateful failover" have become the order of the day.
Notable articles on IP reliability this year:
For more on IP reliability, register for Light Reading's Webinar on the topic, scheduled for January 23, 2003.
5: GFP Packs Up Next-Generation Sonet/SDH
Generic Framing Procedure (GFP) went from being an acronym nobody knew (save some zealots at Lucent Technologies Inc. [NYSE: LU] and Nortel Networks Corp. [NYSE/Toronto: NT]) to a top priority for the multiservice edge of the transport network at most ILECs.
GFP, plain and simple, is just a very elegant way of putting Layer 2 traffic onto Sonet (Synchronous Optical NETwork) and SDH (Synchronous Digital Hierarchy) networks. It encompasses other hot developments such as virtual concatenation, which creates fine-grained connections in a Sonet/SDH network; LCAS (Link capacity adjustment scheme), which dynamically resizes those channels based on operator requests or traffic fluctuations; and RPR, which squeezes the most out of a ring-based network carrying predominantly packet traffic.
GFP also has a bonus feature, called "transparent mode," which supports oddball protocols such as those for storage networking or data center interconnect.
That said, GFP represents something of a missing link in the carrier’s search for the multiservice network: a universal method of framing. Whereas "God boxes" of the past simply put every type of interface possible into a metro box, then added a bunch of switch fabrics to support them, GFP packages up everything nicely and sends it on its way, with true point-and-click operation. And best of all, it’s standardized by the International Telecommunication Union, Standardization Sector (ITU-T), so carriers actually give it the time of day. GFP is hot in 2002, but look for real buildup in 2003, when boxes start shipping en masse.
Notable articles on GFP and next-gen Sonet/SDH this year:
4: The "Multiservice Edge" is Redefined and Reinvigorated
The term "multiservice edge" has been around for ages, but this year it's taken on a new meaning - and new marketing momentum. The old meaning typically refers to transport equipment - notably Sonet boxes capable of handling multiple protocols. In other words, it's more multiprotocol than multiservice.
The new meaning, where most of the activity has been this year, refers to the packet layer, not the transport layer. In this case, multiservice means what it says - multiple services over a packet-based infrastructure.
So what counts as "new" multiservice? Well, Asynchronous Transfer Mode (ATM) always has, so companies like WaveSmith Networks Inc. made our Light Reading's Top Ten Private Companies because they are proving the life left in ATM as a solution for carrying services like Frame Relay and IP-over-ATM.
There’s also the next-gen BRAS (broadband remote access server), which supports multiple services over DSL or cable modems, and which was put out by the likes of Juniper Networks Inc. (Nasdaq: JNPR), Network Equipment Technologies Inc. (net.com) (NYSE: NWK), and others.
Finally, there are a host of edge switch/router vendors that are laying the groundwork for a whole new class of multiservice MPLS systems, using well-heeled draft specs from the Internet Engineering Task Force (IETF) to put any number of Layer 2 services on an MPLS-based packet network.
Multiservice will always be a hard concept to define with clarity, but what we saw happening in 2002 was a continued willingness on the operators' part to look at their legacy infrastructure (from Sonet to DSL to ATM) and figure out how to wring the most service revenue out of it. There are just too many new services out there to build a distinct network layer for each, so multiservice, even though it often gets a bruising from claiming it can be all things to all carriers, is here to stay.
Notable articles on the "new" multiservice edge this year:
3: DSL Proliferates and Evolves
Digital subscriber line (or DSL) technology, though seeming to slow down in the U.S. this year, is huge: a true worldwide phenomenon. Though this began well before 2002, this year is the one in which DSL reached over 30 million subscribers globally.
Every major carrier worldwide has a DSL plan in place with ambitious goals. France Telecom recently said it would wire every town of 5,000 or more. South Korea has wired just about dwelling imaginable, and in the U.S., where ILECs whine about UNE-P (unbundled network elements platform - see RBOCs Get Long Distance Go-Ahead) and whatever else ails them, DSL continues to be a major new source of revenue, if not profit.
The next big step for DSL is business. Already, many small businesses worldwide are using ADSL to connect to the Internet, but this year new classes of DSL technology were rolled out that promise true business-class connectivity. There is ADSL Annex J in Europe, which supports 3-Mbit/s symmetric service, G.SHDSL, which does about the same, various flavors of VDSL that offer speeds up to 50 Mbit/s depending on loop length, and a number of DSL varieties that support the bonding of multiple pairs together to create true full-rate Ethernet connections in the access network.
This is an interesting one: Ethernet over DSL may in some cases spell the demise of DSL as a service and instead relegate it to little more than a “bit pump.” Under this model, carriers would no longer sell DSL services, but rather Ethernet access, which has much more sex appeal.
DSL is also behind one of the other hot trends of 2002, the "Triple Play": a trifecta of voice, video, and data over a single access connection. The independent LECs in the U.S. are interested, as are a number of operators in Europe and Asia. Using VDSL, connections of 20 Mbit/s and beyond can be provisioned to a home, with ample room for three simultaneous cable TV sessions, symmetric data, and voice, all on one supercharged telephone line. The Triple Play has long been sought by telcos facing ambitious MSOs in their backyards, and already there are signs that DSL is paving the way.
Notable articles on DSL this year:
2: 802.11 Wireless Technology Moves to Mass Deployment
IEEE 802.11 wireless LAN technology has been around for a while, but this was the year that everyone actually bought it. WiFi hubs are under $200 and showing up in coffee shops, hotels, airports, and lots of homes and apartments. Enterprise customers get to choose from 802.11b or the higher-speed 802.11a and have embraced it happily.
This raises the question: Where does 3G telephony fit in a world where low-cost WiFi hotspots are popping up on every street corner? Perhaps they'll reach different markets - consumers with 3G, and business users with WiFi.
Notable articles on 802.11 this year:
1: Ethernet Gains Ground in Access Networks
This was a big year for Ethernet standardization, and the feeling around Ethernet can’t be more enthusiastic. The inevitability of Ethernet as the universal interface to the network of the future is now etched in granite, and standards are the chisel. This year we saw standards for 10-Gigabit Ethernet ratified by the IEEE as the 802.3ae standard in June. We also saw the beginnings of 802.3af, which will define how to send power over Ethernet networks. We also saw significant progress in the 802.3ah standard, which defines various physical layers for Ethernet in the First Mile, including fiber and copper.
We also saw the ratification of the RPR (resilient packet ring) standard (IEEE 802.17), which promises to play an important role between Ethernet access networks and Sonet/SDH core telecom infrastructure.
Beyond the official standards, the industry moved forward with multisource agreements (MSAs) for 10-GigE transceivers – with the goal of making 10-GigE interfaces easy to create and plug into a printed circuit board. The only problem is that there are too many MSAs. The first one, XENPAK, now faces competition from XPAK and X2, and an MSA for a serial 10-GigE transceiver – XFP – is also in the works.
The Metro Ethernet Forum (MEF) did its part to help push Ethernet into the service provider network by moving sixteen technical documents to MEF Draft status or beyond in the following four areas: Services, Protocol and Transport, Management, and Architecture. This process is critical in bringing carriers to the necessary comfort level to start rolling out more advanced Ethernet–based services.
Notable articles on Ethernet this year:
10-Gig Ethernet Transceivers
— Scott Clavenna, Director of Research, and Peter Heywood, Founding Editor, Light Reading