The rollout of next-generation broadband services has resulted in an upsurge of interest and sales in IP DSLAMs (digital subscriber line access multiplexers), which has provided an opportunity for new players to enter the market for supplying this type of equipment.

From a service provider perspective, it means going back to basics – reviewing fundamental requirements for delivering triple-play-type services incorporating video, voice, and data, and taking a fresh look at how those requirements can be met by what’s on offer in the equipment marketplace.

Light Reading and its market research division, Heavy Reading, have devised a joint plan for helping service providers undertake this task.

The plan starts with this report, which calls on carriers and their prospective suppliers to help us reach a consensus on three issues:

In order to do this, this report puts forward initial proposals in each these areas and then invites readers to suggest modifications and additions, using the message board attached to this article or by emailing us at [email protected] with “IP DSLAM” included in the subject field.

We will update this report in response to your correspondence in the same way that we have updated each of the Who Makes What reports that have become a regular feature of Light Reading and its associated Websites.

The next phase of this project will be undertaken by Heavy Reading. It will invite all of the vendors of IP DSLAMs to submit details of their products, using a questionnaire based on the list of key characteristics coming out of this report.

The data from the questionnaires will then be processed to create a granular competitive analysis of IP DSLAM products, similar to the ones that Heavy Reading has published in the past, covering Session Border Controllers, Softswitches, and Media Gateways, among others, all produced by Graham Beniston, author of this report. A scoring system will be used based on the weightings given by readers in the interactive poll included on Page 3.

Here’s a hyperlinked summary of the report contents:

Related Light Reading Webinar archives:

Related Heavy Reading Reports:

— Graham Beniston has over 25 years experience in the telecom industry with Marconi (GEC, GPT, and Plessey), where he held a variety of technical, business development, and marketing roles and worked in all narrowband and broadband technologies – most recently as system design authority for triple-play services over DSL networks. His expertise includes IP over ATM, IP VPNs, mobile IP, IP satellite networks, and IP traffic management. He has contributed to Light Reading Webinars and reports on next-generation B-RAS, DSLAMs, and edge routing. Beniston holds an MSc in Data Communications Systems from Loughborough University of Technology and is a member of the Institute of Physics and the IEEE.

What Is an IP DSLAM?

In 2003, Heavy Reading defined an IP DSLAM as one having IP routing capability and other Layer 3 functionality. Limited B-RAS (broadband remote access server)capabilities including PPP (point-to-point protocol) and RFC 1483 termination were also seen as competitive features. (See Next-Generation DSL Equipment: The Path to Profitability.)

But the industry trend to calling any DSLAM with any IP Layer functionality and non-ATM backhaul an IP DSLAM has continued, so we will use this wider description for this project.

This means that we include not only Ethernet DSLAMs with a minimum of IP functionality, but also ATM DSLAMs with minimal or high IP functionality.

This may be contrary to the current perception of IP DSLAMs as being somehow cheaper and having better bandwidth capabilities than ATM DSLAMs. While Ethernet DSLAMs with IP functionality may be cheaper, they are not always denser in terms of number of lines handled and simultaneous bandwidth per line supported.

The main difference between ATM and Ethernet DSLAMs is the interior switching technology. It is quite normal to find ATM DSLAMs with Gigabit Ethernet backhaul capability so that they can compete on capex and opex.

IP DSLAMs can be deployed in the central office (CO) or in remote terminals in the outside plant (OSP). The market and regulatory drivers are making OSP deployment more popular, with environmentally hardened units located in a range of housings, from cabinets to pole mountings or underground enclosures.

In North America, the new trend is to put the IP DSLAM at the serving-area interface (SAI) and limit the subscriber copper drop length to 5,000 feet. CO IP DSLAMs are usually, but not always, housed in large chassis and rack systems. OSP IP DSLAMs are usually smaller and come in a variety of form factors. The defining property of an OSP DSLAM in this study is that it is environmentally hardened for deployment in unconditioned housings.

Why Are They Hot?

Telco TV and video service support
The reason IP DSLAMS are currently hot products is the almost universal decision by operators to deploy TV and video services over DSL. The drivers have been well documented elsewhere.

This decision has major implications for DSLAMs. TV and video streams require much higher bit rates than does general-purpose high-speed Internet access or voice-over-IP services. With each video stream requiring between 1 and 4 Mbit/s, operators are looking to increase the rate of DSL services offered to between 5 and 25 Mbit/s. Since the supportable bit rate of a particular line is dependent on its length, the need to maintain a viable penetration of the market with advanced video services has guaranteed the need to deploy remote DSLAMs in street cabinets. The high bit-rate requirement has also increased the interest in DSL line bonding.

The high-rate video streams will of themselves require higher throughput in the DSLAM. But the potential requirement to decrease normal contention ratios for residential Internet access from around 50:1 to 10:1 or even 1:1 for video-on-demand service puts a much higher importance on the DSLAM throughput.

The need to include broadcast TV in the service set gives an added requirement for multicast support. The minimum support required is IGMP snooping. Higher competitiveness is achieved with a greater IGMP capability, particularly proxy support and, to a lesser extent, multiple version support. For DSLAMs with IP routing capability, IP multicast support is highly desirable. The preferred protocol is PIM-Sparse Mode.

A knock-on effect of the requirement for higher throughput is the cost of B-RAS equipment. There is a growing feeling in operators that putting high bit-rate video services through a centralized B-RAS is not cost effective. One proposed solution, which increases the competitiveness of DSLAMs, is the inclusion in them of a distributed subset of B-RAS capability. As part of this distribution of functionality, at least two alternative architectures are being considered: one with minimal IP functionality at the DSLAM, and one with enhanced IP functionality, including IP routing.

Multiservice access platform (MSAP) requirement
While acknowledging that telco TV is the current white-hot driver for IP DSLAM deployment, there is still a rapidly growing requirement for multiservice support over DSL to increase ARPU, particularly VOIP, including the translation of analogue voice to VOIP at the DSLAM for Class 5 migration.

While we have decided not to include a comparison of MSAPs here, we have listed the products involved. One implication that does cross over into IP DSLAM competitiveness is the requirement to support quality of service (QOS) handling for multiple services. The minimum requirement is support for IEEE 802.1p prioritization and IEEE 802.1Q VLAN support. Much better is the inclusion of IP Diffserv handling and MPLS traffic engineering.

The defining properties of an MSAP are the inclusion of a VOIP media gateway, an FTTx optical line terminal (OLT), and an integrated packet transport for all traffic types. MSAPs can be deployed in CO locations or can be hardened for OSP use. We are not performing a detailed comparison, so all forms are included in the product listings.

OSP deployment
We have explained why DSLAM deployment remote from the CO is growing rapidly. The accommodation for these DSLAMs is quite varied, from small spaces in existing cabinets to environmentally controlled huts and vaults.

Which OSP IP DSLAM sizes should we compare? The only standardized sizes appear to be the 1U pizza box or the cut-down 19-inch CO chassis. Are there other mini-IP DSLAM boxes we should take into account? We certainly think that all OSP IP DSLAMs we compare should be environmentally hardened to differentiate them from redeployed CO chassis units.

A proposed list of key features and performance metrics is given below. We believe that all of these features are relevant to IP DSLAMs providing high-density TV and video-over-DSL services in addition to high-speed Internet access.

In order to compare products using these metrics, Heavy Reading will need to give each one a weighting to reflect its relative importance to the average service provider. The poll below gives readers an opportunity to share their opinions on which characteristics are most important – and modify the weightings we use in the Heavy Reading study.

Please rate the importance of each of the listed features:

{survey 110}Other features?
If we’ve missed something you consider important please tell us on the message board or email [email protected], citing “IP DSLAM” in the subject field.

Where’s Your Firm? Where’s Your Product?

It’s been hard to track every vendor and product because of the explosion of interest in this product area, especially from startups in Asia and Eastern Europe. So if your firm has a product, and either the firm or the product is not mentioned below, let us know. Again, to be listed in the outside plant (OSP) section requires an environmentally hardened unit.

Table 1: IP DSLAM Suppliers

Other?
If you see that we’ve missed any company or products, please tell us on the message board or email [email protected], citing “IP DSLAM” in the subject field.