While we have been talking about Phantom DSL for the past few weeks, wading through our archives and listening to some DSL aficionados at Airties, where we visited this week, we were reminded of Dynamic Spectrum Management (DSM), which back in 2006 was the great white hope for DSL.
The principle behind DSM is the avoidance of crosstalk. In most wired communications this is not a problem, but when a telco takes lines out of an exchange, it automatically bundles them up for easy maintenance and with no shielding on the wires this leads to crosstalk ‘ sideways interference from line to line.
Israeli telecommunications equipment vendor ECI Telecom has had a program in place since 2006 and even got government money to research the subject and quoted us chapter and verse from John Cioffi, at the time Professor of Engineering at Stanford University, and known as the Father (virtually the inventor) of ADSL and also acknowledged as one of the inventors of OFDM.
DSM uses four key concepts which are channel identification, spectrum balancing, vectored transmission and multi-user detection. The idea is to decide which spectrum to use on adjacent lines to limit interference within each twisted pair binder and this requires individual line management. Back in 2006 they were talking about rates of 45 Mbps at distances of over 1,500 feet from an exchange, and a doubling of the speed of any DSL over long-distances, making services possible on 12,000 foot loops.
Now Cioffi is working at US firm Assia just a few weeks ago landed a $20.8 million investment round with cash from Telefonica, AT&T, and Sandalwood Partners to go with previous investors Mingly China Growth Fund, SFR Development, Sofinnova Partners, Stanford University, Swisscom Ventures, and T-Ventures.
What Assia promises is that (eventually) it can get DSL to 100 Mbps over existing phone lines. It does this by interposing servers called element management systems which collect operations and maintenance data on DSL lines on demand. Using this data it can automate DSL service optimization on a per-line basis. Its handful of customers claim dramatic speed and reach improvements, without ever changing a DSL line card. Assia is already working with AT&T in the US to ensure its DSL line speeds.
So all is not doom and gloom for DSL, and it doesn’t need to wait about 3 years for phantom DSL and other technologies to become standards and find their way into chips. This approach can work, and is working, now.
In October Assia launched Expresse 2.1, an update on earlier versions of its DSM management system which claims it can cut problems in a DSL network by 30% to 60% and offer a 40% increase in speed and service reach. It reckons that this system can serve 20 million DSL lines from as few as 4 servers, and brings with it automatic line repair, plus the optimization of all network lines.
When it is working with VDSL2, the company says that it can already support downstream data speeds up to 200 Mbps and can even manage bonded lines. Of course that’s presuming that you were close enough for VDSL2 to work in the first place, at its stated rate of around 100 mbps, which would be some 1,200 feet from a DSLAM.
The company also offers a Power Management module which it says can save up to $500,000 a year in energy costs over a 1 million line network, with no performance penalties.
Of course the upshot of all this is that for the cost of a few servers and some software and a bit of handholding, plus some procedural changes, a telco can prolong its investment in current technology as the same time as including far more ‘reach’ customers, who were previously below the distance threshold for broadband, but also do the same for IPTV customers. Often telcos will only provide IPTV to customers which can get a 10 Mbps throughput on their broadband line, or higher. If such techniques double the speed of the DSL lines, then very likely twice as many are within reach of IPTV, with all the ARPU implications that brings. No wonder these operators were keen to put over $20 million into the pot.