How many 28.8k lines in one T1
Golaq
I was wondering how many 28.8k lines you can get out of a T1 line.
Rick Ellis
Rich Greenberg
In article <19970305210...@ladder01.news.aol.com>,
Golaq <go...@aol.com> wrote:
>I was wondering how many 28.8k lines you can get out of a T1 line.
It depends how you are going to do it.
If the T-1 contains the usual 24 voice channels (64kb each) and you are
using V.34 modems then one modem signal per channel.
If the 28.8 are already digital signals and the T-1 is configured as a
1.5+ megabit pipe, you could multiplex about 50 of the 28.8 streams
onto it. I don't know if off-the-shelf hardware could do this.
--
Rich Greenberg
N6LRT TinselTown, CA, USA Play: ric...@netcom.com 310-649-0238
Pacific time. I speak for myself & my dogs only. VM'er since CP-67
Canines: Val(Chinook,CGC), Red(Husky,(RIP)), Shasta(Husky) Owner:Chinook-L
Joel M Snyder
In article <331e198d....@news.thenet.net>, ktc...@thenet.net (Tim
Keating) writes:
> el...@ftel.net (Rick Ellis) wrote:
>
>>In article <19970305210...@ladder01.news.aol.com>,
>>Golaq <go...@aol.com> wrote:
>>>I was wondering how many 28.8k lines you can get out of a T1 line.
>>
>
> Not correct.. To get 24 you must use in-band signalling... Which drops the
> effective rate down 26.4kb send and receive..
>
> For 28.8 you should use PRI T1, or 23B+1d.. which means only 23 true 28.8
> (up to 33.6) connections can be made.
You're badly confused. A T1 ESF/B8ZS gives 24 perfectly beautiful voice
channels, all of which are capable of running 33.6 using standard V.34
technology. Whether or not the user should jump from CT1 to ISDN
technology (given that they want analog modems to dial in) is a question of
tariff.
jms
Joel M Snyder, 1404 East Lind Road, Tucson, AZ, 85719
Phone: +1 520 324 0494 (voice) +1 520 324 0495 (FAX)
j...@Opus1.COM http://www.opus1.com/jms Opus One
Tim Keating
el...@ftel.net (Rick Ellis) wrote:
>In article <19970305210...@ladder01.news.aol.com>,
>Golaq <go...@aol.com> wrote:
>>I was wondering how many 28.8k lines you can get out of a T1 line.
>
>24
Not correct.. To get 24 you must use in-band signalling... Which drops the
effective rate down 26.4kb send and receive..
For 28.8 you should use PRI T1, or 23B+1d.. which means only 23 true 28.8
(up to 33.6) connections can be made.
Tim K.
Jeremiah Kristal
On Thu, 06 Mar 1997 01:13:19 GMT, ktc...@thenet.net (Tim Keating)
wrote:
>el...@ftel.net (Rick Ellis) wrote:
>>In article <19970305210...@ladder01.news.aol.com>,
>>Golaq <go...@aol.com> wrote:
>>>I was wondering how many 28.8k lines you can get out of a T1 line.
>>24
>Not correct.. To get 24 you must use in-band signalling... Which drops the
> effective rate down 26.4kb send and receive..
I take it you've never heard of B8ZS.
>For 28.8 you should use PRI T1, or 23B+1d.. which means only 23 true 28.8
>(up to 33.6) connections can be made.
Nope, if you're able to get 1 33.6 connection on a T1, you should be
able to get 24. A T1 has 24 discrete 64k channels (on a B8ZS/ESF T1),
and the only advantage of a PRI is that you can get 64k rather than
56k ISDN connections
Jeremiah
danny burstein
while the official answer (see the folk quoted below my comments) is (kind
of) 24, the question needs to be more precise.
If you're talking about regular phone lines which just happen to have v.34
(28k/33k) modems on them, yes the answer is 24.
If you're looking at comparing how many 28k lines does it take to equal
the data rate of a T1, then when you work out the math (T1=1.54 meg)
you'll see a number more like 55. Note that this refers to a fully loaded
28k line (see below). (let's leave out latency and other overhead)
If you're thinking of 'how many 28k dial-up customers can a T1 support'
you'll get to a much a higher number. Remember that the folk calling in
are _not_ all simoultaneously saturating their 28k connections, so you can
de-rate the figures. Actual percentage will vary a great deal depending on
custmer mix, what they're doing, and time of day.
So there are multiple answers to your question, ranging from 24 to (very
roughly) 150.
>>el...@ftel.net (Rick Ellis) wrote:
>>>24
>Jeremiah
--
_____________________________________________________
Knowledge may be power, but communications is the key
dan...@panix.com
Rmcg
With the right T1 equipment you could get 52 to 53 28.8's from 1 t1.
1.544m in a t1. devided by 28.8 gives you 53 plus 28.8's, but there is
always a little overhead in a t1 mux.
Golaq <go...@aol.com> wrote in article
<19970305210...@ladder01.news.aol.com>...
Thomas Laman
Jeremiah Kristal wrote:
>
> On Thu, 06 Mar 1997 01:13:19 GMT, ktc...@thenet.net (Tim Keating)
> wrote:
>
> >el...@ftel.net (Rick Ellis) wrote:
>
> >>In article <19970305210...@ladder01.news.aol.com>,
> >>Golaq <go...@aol.com> wrote:
>
>
> >Not correct.. To get 24 you must use in-band signalling... Which drops the
> > effective rate down 26.4kb send and receive..
>
> I take it you've never heard of B8ZS.
>
> >For 28.8 you should use PRI T1, or 23B+1d.. which means only 23 true 28.8
> >(up to 33.6) connections can be made.
>
> Nope, if you're able to get 1 33.6 connection on a T1, you should be
> able to get 24. A T1 has 24 discrete 64k channels (on a B8ZS/ESF T1),
> and the only advantage of a PRI is that you can get 64k rather than
> 56k ISDN connections
>
> Jeremiah
On a B8ZS/ESF T1, where does the signalling information for each channel
come from?
It either comes from the least significant bit of each channel or one of
the 24 channels must be reserved for signalling
information for the remaining 23 channels.
B8ZS allows for 64Kbps transmission without violating the 1's density
requirements of the T1 span.
The first answer was correct.
At 28.8Kbps, you need to have a T1 channel that is CLEAR which means
that all of the channel bandwidth is available for the user (full 64K),
but to do this, B8ZS line coding is required on the T1 span. Since the
channel itself does not contain the signalling information, it must be
somewhere - on one of the T1 channels which is dedicated to signalling.
Due to this fact, only 23 channels would be available for the users.
Tom
Tim Keating
Thomas Laman <tla...@vivanet.com> wrote:
>Jeremiah Kristal wrote:
>>
>> On Thu, 06 Mar 1997 01:13:19 GMT, ktc...@thenet.net (Tim Keating)
>> wrote:
>>
>> >el...@ftel.net (Rick Ellis) wrote:
>>
>> >>In article <19970305210...@ladder01.news.aol.com>,
>> >>Golaq <go...@aol.com> wrote:
>> >>>I was wondering how many 28.8k lines you can get out of a T1 line.
>>
>> >>24
>>
>> >Not correct.. To get 24 you must use in-band signalling... Which drops the
>> > effective rate down 26.4kb send and receive..
>>
>> I take it you've never heard of B8ZS.
>>
>> >For 28.8 you should use PRI T1, or 23B+1d.. which means only 23 true 28.8
>> >(up to 33.6) connections can be made.
>>
>> Nope, if you're able to get 1 33.6 connection on a T1, you should be
>> able to get 24. A T1 has 24 discrete 64k channels (on a B8ZS/ESF T1),
>> and the only advantage of a PRI is that you can get 64k rather than
>> 56k ISDN connections
I've seen in-band signalling at work on T1's for two years, now.. My old ISP
(which was located was 1.5 miles from my location, using the same central
office) used it.
It resulted in very unreliable(dropped every 30 minutes or less) connections
maxing out at 26.4.
I switch to a new ISP(60 miles away) using PRI T1's.. I get rock solid 31.6
connections all the time.
>On a B8ZS/ESF T1, where does the signalling information for each channel
>come from?
>
>It either comes from the least significant bit of each channel or one of
>the 24 channels must be reserved for signalling
>information for the remaining 23 channels.
>
>B8ZS allows for 64Kbps transmission without violating the 1's density
>requirements of the T1 span.
>
>The first answer was correct.
The question was how many 28.8k lines in one T1. Not two, Not Three, Only ONE.
Thus answer 23 lines running 28.8k or greater analog modem connections is
the correct answer.
>
>At 28.8Kbps, you need to have a T1 channel that is CLEAR which means
>that all of the channel bandwidth is available for the user (full 64K),
>but to do this, B8ZS line coding is required on the T1 span. Since the
>channel itself does not contain the signalling information, it must be
>somewhere - on one of the T1 channels which is dedicated to signalling.
>Due to this fact, only 23 channels would be available for the users.
^^^^^
I see we agree.. Thank you..
Tim Keating.
Joel M Snyder
I don't know where you people learned about telecom, but it's clearly not
by understanding how it works or actually doing it.
A T1 with B8ZS/ESF has 24 channels. Not 23. Each of those channels
is receiving a digitized version of the analog signal. If you back that
into a channel bank, break it out to 24 analog lines, and plug each one
into a 33.6 modem, then all 24 channels will support 33.6 connections.
Lest you continue to blather in ignorance, let me show you the output
from our PM3, currently connected with a channelized T1, ESF/B8ZS, and
loop start signalling. All 24 channels are connected; most of them
are at 28.8 or better; at least a quarter of them are at 33.6.
I guess now I'm beginning to understand why consultants get such a bad
name in this business...
jms
Joel M Snyder, 1404 East Lind Road, Tucson, AZ, 85719
Phone: +1 520 324 0494 (voice) +1 520 324 0495 (FAX)
j...@Opus1.COM http://www.opus1.com/jms Opus One
Port Speed Mdm Host Type Status Input Output Pend
---- ----- --- ---------------- ------ ------------- ---------- ---------- ----
S0 14400 M7 ptp0 Netwrk ESTABLISHED 11937217 86132067 4404
S1 26400 M4 ptp1 Netwrk ESTABLISHED 25578751 189690090 0
S2 28800 M8 ptp2 Netwrk ESTABLISHED 25862713 188156392 0
S3 28800 M10 ptp3 Netwrk ESTABLISHED 14461561 178008336 67
S4 26400 M1 ptp4 Netwrk ESTABLISHED 10763321 144302709 0
S5 33600 M15 ptp5 Netwrk ESTABLISHED 12701520 103578713 1159
S6 31200 M13 ptp6 Netwrk ESTABLISHED 8011121 73820025 0
S7 33600 M16 ptp7 Netwrk ESTABLISHED 8992861 77973686 0
S8 28800 M12 ptp8 Netwrk ESTABLISHED 5644915 56721439 0
S9 33600 M18 ptp9 Netwrk ESTABLISHED 2703797 25190454 213
S10 28800 M21 ptp10 Netwrk ESTABLISHED 1481931 28767069 0
S11 31200 M23 ptp11 Netwrk ESTABLISHED 676788 6157771 0
S12 14400 M2 ptp12 Netwrk ESTABLISHED 59014808 44813908 0
S13 33600 M0 ptp13 Netwrk ESTABLISHED 28776204 53924275 0
S14 9600 M6 ptp14 Netwrk ESTABLISHED 6115531 27413357 0
S15 33600 M19 ptp15 Netwrk ESTABLISHED 3121509 16000556 12
S16 26400 M5 ptp16 Netwrk ESTABLISHED 23354029 63158016 0
S17 28800 M11 ptp17 Netwrk ESTABLISHED 13319954 18790968 0
S18 33600 M3 ptp18 Netwrk ESTABLISHED 23510916 207506944 0
S19 33600 M9 ptp19 Netwrk ESTABLISHED 16126264 53598938 0
S20 31200 M20 ptp20 Netwrk ESTABLISHED 772155 213963 1284
S21 33600 M14 ptp21 Netwrk ESTABLISHED 14554059 38640558 0
S22 33600 M17 ptp22 Netwrk ESTABLISHED 138154265 76930119 0
S23 33600 M22 ptp23 Netwrk ESTABLISHED 5407296 15267074 0
Michael Dillon
danny burstein wrote:
> If you're looking at comparing how many 28k lines does it take to equal
> the data rate of a T1, then when you work out the math (T1=1.54 meg)
> you'll see a number more like 55. Note that this refers to a fully loaded
> 28k line (see below). (let's leave out latency and other overhead)
That's not "math". That's "arithmetic". If you want to work out the math
you need to know about stuff like queuing theory and Poisson
distributions.
> If you're thinking of 'how many 28k dial-up customers can a T1 support'
> you'll get to a much a higher number. Remember that the folk calling in
> are _not_ all simoultaneously saturating their 28k connections, so you can
> de-rate the figures. Actual percentage will vary a great deal depending on
> custmer mix, what they're doing, and time of day.
>
> So there are multiple answers to your question, ranging from 24 to (very
> roughly) 150.
ISP's see closer to 200 for this number from empirical experience and
that
is with a T1 line that is carrying other traffic such as an incoming
newsfeed.
The real answer to this question is that it depends on your customer
base and their usage patterns and it is something that needs to be
monitored.
However, if the person wanted to use a T1 to bring dialin lines from a
remote location to a local modem bank then the answer is simple. There
are 24 channels in a DS1 and no magical way to compress them and still
support 28,800 bps modem connections.
--
Michael Dillon - Internet & ISP Consulting
Memra Software Inc. - Fax: +1-250-546-3049
http://www.memra.com - E-mail: mic...@memra.com
Floyd Davidson
Tim Keating <ktc...@thenet.net> wrote:
>Thomas Laman <tla...@vivanet.com> wrote:
>>Jeremiah Kristal wrote:
>>>ktc...@thenet.net (Tim Keating) wrote:
>>>>el...@ftel.net (Rick Ellis) wrote:
>>>>>Golaq <go...@aol.com> wrote:
>>>>>>I was wondering how many 28.8k lines you can get out of a T1 line.
>>>
>>>>>24
>>>
>>> >Not correct.. To get 24 you must use in-band signalling... Which drops the
>>> > effective rate down 26.4kb send and receive..
Probably the _vast_ majority of DS1 level equipment in use today
still has robbed bit signaling enabled, and yet is still
providing 28.8Kbps (and 33.6Kbps) connections for v.34 modems.
It works just fine, with a couple caveats.
>>> I take it you've never heard of B8ZS.
B8ZS is unrelated to whether robbed bit signaling is used or not.
>>> >For 28.8 you should use PRI T1, or 23B+1d.. which means only 23 true 28.8
>>> >(up to 33.6) connections can be made.
>>>
>>> Nope, if you're able to get 1 33.6 connection on a T1, you should be
>>> able to get 24. A T1 has 24 discrete 64k channels (on a B8ZS/ESF T1),
>>> and the only advantage of a PRI is that you can get 64k rather than
>>> 56k ISDN connections
Typical interswitch trunking uses 24 channels per DS1, may or
may not be B8ZS, and even when SS7 signaling is used will
commonly have robbed bit signaling enabled (if the facilities
are old enough that it can't be optioned off).
For example, on many DMS-100 switching systems the DS-1
interface cards are old enough that AMI/SF and robbed bit
signaling are the only possible options.
>>On a B8ZS/ESF T1, where does the signalling information for each channel
>>come from?
>>
>>It either comes from the least significant bit of each channel or one of
>>the 24 channels must be reserved for signalling
>>information for the remaining 23 channels.
>>
>>B8ZS allows for 64Kbps transmission without violating the 1's density
>>requirements of the T1 span.
>>
All of the above is correct.
>
> The question was how many 28.8k lines in one T1. Not two, Not Three, Only ONE.
> Thus answer 23 lines running 28.8k or greater analog modem connections is
>
> the correct answer.
24 is common.
>>At 28.8Kbps, you need to have a T1 channel that is CLEAR which means
>>that all of the channel bandwidth is available for the user (full 64K),
Probably a majority of 28.8Kbps connections are made over
equipment that is not 64Kbps clean. In the best case that only
amounts to the least significant bit in every 6th frame being
wrong half of the time. In the worst case it is each and every
frame being wrong half of the time (assuming a 50-50 ratio of
1's and 0's for the desired PCM signal, it will be wrong 50% of
the time if it is locked to either state).
That certainly limits ISDN or other data use of the channel to
56Kbps, but the distortion it causes for v.34 modems is just a
reduction in the Signal-to-Noise Ratio from a maximum of about
37 dB to about 31 dB in the absolute worst case (which probably
will limit connections to 26.4Kbps). In most cases (one frame
out of 3 or 6 being robbed) it probably results in a maximum
SNR of 34 or 35 dB.
>>but to do this, B8ZS line coding is required on the T1 span. Since the
>>channel itself does not contain the signalling information, it must be
B8ZS is not required on any channelize DS1 where all channels
are connected to a codec (i.e., a standard POTS line), because
the codec never generates a 0 byte value, hence there will never
be insufficient 1's density. V.34 modem connections, unlike
ISDN in data mode, have no need for B8ZS. That is why virtually
all DS-1 interface cards installed for switch trunks on
DMS-100 switches until about 1990 or so did not even have an
option for B8ZS (or ESF for that matter). And part of the "huge
cost" to implement ISDN that the telco's complain about is
because they love to toss in the cost of replacing all those old
DS-1 interface cards to ramp up the price. (I'm not familiar
enough with switching systems other than the DMS family to know
if the same applies or not.)
>>somewhere - on one of the T1 channels which is dedicated to signalling.
>>Due to this fact, only 23 channels would be available for the users.
> ^^^^^
> I see we agree.. Thank you..
Nahhh... :-)
Floyd
--
Floyd L. Davidson Salcha, Alaska fl...@tanana.polarnet.com
Rick Ellis
In article <331F42...@vivanet.com>,
Thomas Laman <tla...@vivanet.com> wrote:
>On a B8ZS/ESF T1, where does the signalling information for each channel
>come from?
>
>It either comes from the least significant bit of each channel or one of
>the 24 channels must be reserved for signalling
>information for the remaining 23 channels.
>
>B8ZS allows for 64Kbps transmission without violating the 1's density
>requirements of the T1 span.
>
>The first answer was correct.
>
>At 28.8Kbps, you need to have a T1 channel that is CLEAR which means
>that all of the channel bandwidth is available for the user (full 64K),
>but to do this, B8ZS line coding is required on the T1 span. Since the
>channel itself does not contain the signalling information, it must be
>somewhere - on one of the T1 channels which is dedicated to signalling.
>Due to this fact, only 23 channels would be available for the users.
You do NOT need a 64k clear channel to get a 28.8k modem connection.
I am quite certain of this because I connect to a modem via a robbed
bit signaled T-1 channel quite often.
Peter Pflaum
I am working on a news search page - http://www.wiredbrain.com/ then
sample.htm Let me know if you find this useful ?
bobm
Golaq wrote:
>
> I was wondering how many 28.8k lines you can get out of a T1 line.
Tim Keating
I gather your utility on logs initial Connect speeds??
Give the modem's a few minutes and they'll drop the connect speeds as the
errors add up..
Tim K.
Tim Keating
ktc...@thenet.net (Tim Keating) wrote:
>j...@tennis.opus1.com (Joel M Snyder) wrote:
>
>>I don't know where you people learned about telecom, but it's clearly not
>>by understanding how it works or actually doing it.
>>
>>A T1 with B8ZS/ESF has 24 channels. Not 23. Each of those channels
>>is receiving a digitized version of the analog signal. If you back that
>>into a channel bank, break it out to 24 analog lines, and plug each one
>>into a 33.6 modem, then all 24 channels will support 33.6 connections.
Hmm, I notice from your table.. That the Modem numbers don't line up with
the Port numbers.. Perhaps they're not all connected to the T1 as you
claim???
Note: It's fairly difficult to query modem speeds once the PPP/Slip connections
are established. On most modems you'll never know that link rate has changed
after the initial connection.
Also don't forget to also log the transmit and recieve speeds.
Tim K.
Victor Escobar
On 6 Mar 97 15:39:41 GMT, "Rmcg" <95m...@apc.net> wrote:
>With the right T1 equipment you could get 52 to 53 28.8's from 1 t1.
>1.544m in a t1. devided by 28.8 gives you 53 plus 28.8's, but there is
>always a little overhead in a t1 mux.
>
>Golaq <go...@aol.com> wrote in article
><19970305210...@ladder01.news.aol.com>...
>> I was wondering how many 28.8k lines you can get out of a T1 line.
>>
Since a T1=24B+D, wouldn't it be 1.536 + an 8Kbps delta channel?
Therefore the arithmetic should take into account the 1.536 figure,
not the 1.544.
-----------------------
Victor Escobar |
Internet Consultant |
-----------------------
John R. Thomas
go...@aol.com (Golaq) wrote:
>I was wondering how many 28.8k lines you can get out of a T1 line.
The correct answer is 53.61111111 at 28.8 and 45.955238095238 at 33.6
regardless of the line coding B8ZS or AMI or Clear channel or PRI.
The problem is simple, data over the voice network is limited to 24.
Now you can sleep at night.
richard neveau
Rmcg wrote:
>
> With the right T1 equipment you could get 52 to 53 28.8's from 1 t1.
> 1.544m in a t1. devided by 28.8 gives you 53 plus 28.8's, but there is
> always a little overhead in a t1 mux.
>
> Golaq <go...@aol.com> wrote in article
> <19970305210...@ladder01.news.aol.com>...
> >
Hey what about if you convert the modem data to packets first and
then sent it !! you could use stat muxing to squeeze even more
user data onto a T1.
just baiting the trap....
If you are asking for business reasons see the data collected
by Boardwatch Magazine for industry averages for modem to
T1/"Uplink bandwidth" ratios (for the ISPs that reported).
The normal ratio was about 14.4 Kbps of uplink bandwidth per
modem port connected. this would be about 100 modems per T1.
see: http://www.boardwatch.com/mag/96/nov/bwm1.htm
for more details.
of course the other important ratio is customers per modem
dialup line as any AOLer will be able to testify to !!
Moe Kunkle
bobm wrote:
>
> Golaq wrote:
> >
> > I was wondering how many 28.8k lines you can get out of a T1 line.
This begs the question: How many techs does it take to change a light
bulb?
Moe Kunkle
Luis A. Rodriguez
In article <5fqevf$d...@news1.fast.net>, big...@fast.net (John R. Thomas) wrote:
> go...@aol.com (Golaq) wrote:
>
> >I was wondering how many 28.8k lines you can get out of a T1 line.
>
> The correct answer is 53.61111111 at 28.8 and 45.955238095238 at 33.6
> regardless of the line coding B8ZS or AMI or Clear channel or PRI.
> The problem is simple, data over the voice network is limited to 24.
> Now you can sleep at night.
The correct answer depends on which side of the network you are.
If you are dialing into a communications server you can only connect 24 modems
because a T1 has 24 voice channels.
If you are routing traffic from two servers thru a T1 then it is equivalent to
53.61 28.8 modems or 45.95 33.6 modems.
--
Luis A. Rodriguez
lro...@prtc.net
Floyd Davidson
In article <33238A...@ro.com>, Moe Kunkle <m...@ro.com> wrote:
>bobm wrote:
>>
>> Golaq wrote:
>> >
>> > I was wondering how many 28.8k lines you can get out of a T1 line.
>
>This begs the question: How many techs does it take to change a light
>bulb?
>
>Moe Kunkle
That depends... what is it you want the light bulb changed into?
jim ford
That is a question with many answers.
You can conceivably put as many as you like for a shared low bandwidth
situation, a la X.25.
For end to end analog connection you could break each channel down so
that you have 48 connections with full throughput.
In an analog to digital situation, the number still allowing for 28.8
continuous throughput would is 53.
Jim Ford
Christopher W. (Chris) Boone
jim ford wrote:
>
> For end to end analog connection you could break each channel down so
> that you have 48 connections with full throughput.
A T1 is only good for 24 channels..both ways....not 48!
Me Myself and I
: > For end to end analog connection you could break each channel down so
Personally, I find the repeat postings of this question as more a
trolling attempt than anything else.
The proper question should really be either:
a) How many POTS circuits on a T1
or
b) How much bandwidth does a T1 have, and how much bandwidth does a
typical 28.8 modem user eat while poking around, web browsing etc...
anyway...
WRAPIEN
Christopher W. (Chris) Boone wrote:
>
> A T1 is only good for 24 channels..both ways....not 48!
>
If you use Voice Compression, you can double the number of channels for
analog.
Using ADPCM, you can compress 64kps DSO channels to 32kps channels.
ADPCM is a
compression algorithm that takes the 8-bit PCM samples and converts them
to 4-bit
PCM samples. The compressed DSOs are combined in pairs to form
individual 64kps
DSOs.
Bill
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
I do not speak for DSC
It is not in my job description and I don't get paid for it so I refuse
to
speak for them!
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Jeffrey Rhodes
WRAPIEN wrote:
>
> Christopher W. (Chris) Boone wrote:
> >
> > A T1 is only good for 24 channels..both ways....not 48!
> >
>
> If you use Voice Compression, you can double the number of channels for
> analog.
>
> Using ADPCM, you can compress 64kps DSO channels to 32kps channels.
> ADPCM is a
> compression algorithm that takes the 8-bit PCM samples and converts them
> to 4-bit
> PCM samples. The compressed DSOs are combined in pairs to form
> individual 64kps
> DSOs.
>
> Bill
>
What kind of call control is used on a T-1 that transports ADPCM
calls? SS7 can control only a full T-1 channel, not sub-channels.
When a V.34 modem call is forced to route into this ADPCM
compression, is it still able to establish a 28,800 bps
connection to the called V.34 modem? I don't think so.
Jeffrey Rhodes at jeffrey...@attws.com
Al Varney
In article <335811...@eng.claircom.com>,
Jeffrey Rhodes <jrh...@eng.claircom.com> wrote:
>WRAPIEN wrote:
>> Christopher W. (Chris) Boone wrote:
>> >
>> > A T1 is only good for 24 channels..both ways....not 48!
>>
>> If you use Voice Compression, you can double the number of channels for
>> analog.
>>
>> Using ADPCM, you can compress 64kps DSO channels to 32kps channels.
>> ADPCM is a
>> compression algorithm that takes the 8-bit PCM samples and converts them
>> to 4-bit
>> PCM samples. The compressed DSOs are combined in pairs to form
>> individual 64kps
>> DSOs.
>
>calls? SS7 can control only a full T-1 channel, not sub-channels.
I think you meant "SS7 can control only DS0 channels, or multiples
of DS0 channels, not DS0-sub-channels".
As far as I know, there are no digital switches supporting switching
of sub-DS0-rate encodings such as 32K ADPCM. To switch such voice paths,
the bit stream must be re-converted to a standard DS0 signal. Thus,
ADPCM is invisible to switches and to SS7. In effect, ADPCM is useful
for point-to-point bandwidth conservation. For example, for trans-ocean
applications, for private networks leasing transmission facilities or in
radio-spectrum applications such as cellular mobile air interfaces.
>When a V.34 modem call is forced to route into this ADPCM
>compression, is it still able to establish a 28,800 bps
>connection to the called V.34 modem? I don't think so.
ADPCM compresses VOICE, not voice-band data, so such channels will
not carry everything that fits in a 64K DS0 channel. 32K (4-bit) ADPCM
will handle 4800 b/s modems, but 40K (5-bit) ADPCM is needed to carry
9600 b/s and could handle 14.4 Kb/s is no other impairments exist.
In trans-ocean and satellite applications, it is thought the encoding
equipment could dynamically switch between 2, 3, 4 and 5-bit ADPCM
sub-channels as needed, so that FAX and data applications would not have
to use separate call routing paths. In private network applications,
ADPCM is usually applied to a portion of a DS1 (say, 12 channels), with
the remainder reserved for digital data (say, 384Kbps) and high-quality
voice/voice-band data. This provides economical PBX-to-PBX interfaces
for long distances. During low voice-calling periods, some ADPCM channels
can be re-mapped to data channels, providing higher digital capacity
during such periods.
In the cellular air-interface, you're pretty much stuck with the
bandwidth allotted, unless some cellular packet mechanism is used.
Al Varney - just my opinion
richard neveau
Jeffrey Rhodes wrote:
>
> When a V.34 modem call is forced to route into this ADPCM
> compression, is it still able to establish a 28,800 bps
> connection to the called V.34 modem? I don't think so.
>
Oh yea of little faith. It is called dynamic bandwidth
allocation. You give full 64Kbps to 'data calls' and compress
the hell out of any voice. Silence suppression helps too.
Lot's-o-DSP's monitoring for type of call.
You do need a good voice to data mix or the scheme goes to hell.
Also ADPCM may not be the best compression algorithm but that
has nothing to do with my answer.
It is NOT impossible, just a pain.