SDSL reverse engineering presented at REcon 2010

by Michael Sokolov,

URL : http://www.recon.cx/2010/slides/sdsl_journey.pdf

Summary : SDSL is a trailing-edge telecom technology that was originally intended to fill the gap between consumer ADSL and business T1/FT1 services. When I started working with SDSL in 2004, I had chosen it because it allowed me to remain a "business" customer (as opposed to consumer/ residential), have symmetric up & down speeds (I would rather have a low symmetric speed than high asymmetric), go faster than ISDN, yet pay only $150 to $180 per month instead of upwards of $500 for a T1.
Over the past 5 years I have successfully carried out a project which allows low-speed SDSL (from 160 kbps to T1 speeds) to be used as a still-available replacement for ARPANET and for the old 1980s-style Internet, for those who miss the latter. More specifically, I have developed a way to use SDSL with traditional 1980s routers of the late ARPANET / early Internet era.
ARPANET and early Internet ran over 56 kbps DDS and other leased lines; a line of that type is a pipe that carries a synchronous serial bit stream. The WAN interfaces on the classic 1980s / early 1990s routers are thus designed to attach to synchronous serial bit stream media. As it happens, SDSL is also a synchronous serial bit stream, but because it came about in the days when traditional WAN interfaces were going out of fashion in favor of Ethernet, CPE that would allow SDSL to be used in the old-fashioned manner was never made widely available.
When I started working with SDSL in 2004, it was severely hobbled by the fact that the only type of CPE available for it were Ethernet-presenting DSL "modems" and routers much like those for consumer ADSL. Not being able to obtain a non-Ethernet CSU/DSU type of CPE device for SDSL, I had set out to design and build one myself, and 5 years later I have scored a complete & total success. The challenge was further complicated by the fact that SDSL/2B1Q was never a real standard, only somewhat of a pseudostandard with a variety of incompatible proprietary flavors.
In this talk I will share the highlights of my journey which has brought me to the present state of having a CSU/DSU-like device which attaches SDSL to a 1980s/90s router's non-Ethernet WAN interface. This journey included social engineering ventures with several legacy SDSL infrastructure vendors, brute force cracking of an encrypted ZIP with SDSL transceiver chip control software source code, and lots of hardware, firmware and wire protocol reverse engineering.
Related project website: http://ifctfvax.Harhan.ORG/OpenWAN/