Thursday, June 08, 2006

ADSL2 Performance Reports: "Good Interim Solution"


This report from a broadband networking forum is interesting. It discusses achievable bandwidths with "next-generation DSL": ADSL2. This engineer calls it a "Good Interim Solution". The question for AT&T (and the other telcos) is this: will streaming HDTV over next-gen, DSL-based IP pipes suffice? Or we destined for another series of broken promises (e.g., ISDN, fiber rollouts to millions of homes, etc.)?

Based upon the telcos' track records, I think we can all surmise the answer to that question.

I can shed some light on the capabilities of newer DSL technologies. I am the network engineer for a small independent telephone company, and we are currently in the process of testing IP video over our ADSL2+ network. We have been pleasantly surprised by the speeds achieved with ADSL2+. Examples of real world speeds:

5k feet= 26400 down / 1408 up (ADSL2+)
5k feet= 23872 down / 2624 up (ADSL2+ Annex M)

14k feet= 10336 down / 960 up (ADSL2+)
14k feet= 8540 down / 1520 up (ADSL2+ Annex M)

17500 feet = 7240 down / 768 up (ADSL2+)

Loop bonding will double these speeds if needed. We are just beginning the video testing, but we have been assured that SD will consume 1.5-2.5 Mbps and HD will consume 6-9 Mbps using MPEG-4.

Let's do the math. My household might need three HDTV streams running simultaneously: me watching Fox News, my wife watching e-TV, and my kids watching MTV American Idol. Let's do the math: three times 8 megabits-per-second (what I've been told the average HD stream requires) = 24 Mbps. Oopsie. Looks like someone's going to have problems deploying this in the real world. In other situations, I'd be surprised. But seeing who is behind these deployments, I'll just suppress a giggle.

Broadband Reports Forums: Good Interim Solution

No comments: