Kitz ADSL Broadband Information
adsl spacer  
Support this site
Home Broadband ISPs Tech Routers Wiki Forum
 
     
   Compare ISP   Rate your ISP
   Glossary   Glossary
 
Please login or register.

Login with username, password and session length
Advanced search  

News:

Pages: [1] 2

Author Topic: VDSL modemís protocol stack - ATM?  (Read 5519 times)

Weaver

  • Addicted Kitizen
  • *****
  • Posts: 7455
  • Retd sw dev; A&A; 4 ◊ 7km ADSL2; IPv6; Firebrick
VDSL modemís protocol stack - ATM?
« on: September 14, 2016, 05:48:45 PM »

Does a VDSL modem still speak ATM to the cab? If not I wonder what the protocol stack is. Without ATM it will be a good bit more efficient at the same sync rate compared with ADSL will it not?
Logged

burakkucat

  • Global Moderator
  • Senior Kitizen
  • *
  • Posts: 26918
  • Over the Rainbow Bridge
    • The ELRepo Project
Re: VDSL modemís protocol stack - ATM?
« Reply #1 on: September 14, 2016, 06:25:52 PM »

A modem configured for a G.993.2 (VDSL2) service will use PTM (packet transfer mode).
Logged
:cat:  100% Linux and, previously, Unix. Co-founder of the ELRepo Project.

Please consider making a donation to support the running of this site.

ejs

  • Kitizen
  • ****
  • Posts: 1971
Re: VDSL modemís protocol stack - ATM?
« Reply #2 on: September 14, 2016, 06:38:59 PM »

Yes, it uses PTM, although the option to use ATM is still a part of G.993.2.

PTM is part of G.992.3 (ADSL2), but I don't know if any ISPs actually use it on ADSL2/2+.
Logged

kitz

  • Administrator
  • Senior Kitizen
  • *
  • Posts: 32108
  • Trinity: Most guys do.
    • http://www.kitz.co.uk
Re: VDSL modemís protocol stack - ATM?
« Reply #3 on: September 16, 2016, 10:52:27 AM »

I believe somewhere around 12Mbps is when ATM v PTM/ethernet is the crux point.   
The smaller ATM packet sizes were deemed more efficient for ADSL1 to avoid serialisation delay.
Logged
Please do not PM me with queries for broadband help as I may not be able to respond.
-----
How to get your router line stats :: ADSL Exchange Checker

ejs

  • Kitizen
  • ****
  • Posts: 1971
Re: VDSL modemís protocol stack - ATM?
« Reply #4 on: September 16, 2016, 03:54:47 PM »

I think we've had this discussion before, but how can there possibly be any crux point?

ATM = 5 bytes overhead (ATM cell header) + 48 bytes data
PTM = 1 byte overhead (sync octet) + 64 bytes data

It's the ATM that is largely responsible for the ADSL IP Profile being 88.2%, and the PTM for the VDSL2 IP Profile being about 96%.
Logged

Weaver

  • Addicted Kitizen
  • *****
  • Posts: 7455
  • Retd sw dev; A&A; 4 ◊ 7km ADSL2; IPv6; Firebrick
Re: VDSL modemís protocol stack - ATM?
« Reply #5 on: September 16, 2016, 04:57:17 PM »

I believe (hope this is correct) that PTM is implemented using HDLC, which involves escaping (aka byte-stuffing). This could double the size of the data in a pathological frame comprised entirely of byte values that need to be escaped. (Bytes 0x7d or 0x7e if memory serves, it's been a long time.)
Logged

ejs

  • Kitizen
  • ****
  • Posts: 1971
Re: VDSL modemís protocol stack - ATM?
« Reply #6 on: September 16, 2016, 05:25:27 PM »

Quote from: G.992.3
K.3.8.2 Functionality
Two optional packet encapsulation methods are defined:
ē HDLC encapsulation, as defined in clause H.4 of [ITU-T G.993.1];
ē 64/65-octet encapsulation, as defined in Annex N.
Logged

Weaver

  • Addicted Kitizen
  • *****
  • Posts: 7455
  • Retd sw dev; A&A; 4 ◊ 7km ADSL2; IPv6; Firebrick
Re: VDSL modemís protocol stack - ATM?
« Reply #7 on: September 16, 2016, 05:29:03 PM »

@ejs - is it the latter that is used by BT? I haven't yet read annex N
Logged

ejs

  • Kitizen
  • ****
  • Posts: 1971
Re: VDSL modemís protocol stack - ATM?
« Reply #8 on: September 16, 2016, 06:04:33 PM »

Yes, by VDSL2, not only BT.
Logged

NewtronStar

  • Kitizen
  • ****
  • Posts: 4817
Re: VDSL modemís protocol stack - ATM?
« Reply #9 on: September 16, 2016, 10:12:34 PM »

Is this relevant from a cut & paste ?

PPPoE vs PPPoA

Many internet users are often plagued with the decision of what kind of protocol to use with their connection. There are many questions, experimentation, and flip-flopping between PPPoE and PPPoA. Unfortunately, many of the confusions hover above these two protocols.

Both of these protocols give internet service providers (ISP) the capability to rollout broadband use. Using these protocols gives relative security as it obliges the end-user to verify or authenticate to a server before having network access.

PPPoE and PPPoA are mainly used with Digital Subscriber Line (DSL). Both protocols provide flexibility in billing which is very advantageous to providers. Moreover, the use of network resources can be monitored easily and along with this feature, troubleshooting and managing network usage is less problematic.

PPPoE is an abbreviation Point-to-Point Protocol over Ethernet and it is a kind of networking protocol that encapsulates PPP frames inside the Ethernet frames. Basically, PPPoE is configured as a point to point connection between two Ethernet ports.

This protocol is normally used in ISPís lower packages as PPPoE is often utilized to function in lower bandwidth throughput rate. PPPoE is perhaps the most common protocol used when connecting to the internet.

Point-to-Point Protocol over ATM (PPPoA) is also a network protocol but this time, it is for encapsulating frames inside AAL5 or ATM Adaption Layer 5. ATM means Asynchronous Transfer Mode, a kind of switching using time division multiplexing in an asynchronous manner.

In the enterprise packages, PPPoA is typically the protocol of choice. It will require a Static IP configured straight into the hardware (modem). PPPoA is said to have less overheads compared to PPPoE therefore, the former is slightly faster than the latter. However, for the end-user, the difference in speed is almost negligible.

Modems that support PPPoA use Asynchronous Transfer Mode ó which uses very small, fixed-length packets, in contrast to Ethernet, which uses relatively large, variable-length packets ó to do whatever it needs to do.

Summary:

1. PPPoE means Point-to-Point Protocol over Ethernet while PPPoA means Point-to-Point Protocol over ATM.

2. PPPoE encapsulates PPP frames inside the Ethernet frames while PPPoA encapsulates frames inside AAAL5.

3. PPPoE is often used in lower packages while PPPoA is often used in enterprise packages.

4. PPPoE is more commonly used than PPPoA.

5. PPPoA has slightly less overhead than PPPoE therefore, it is a bit faster.





Logged

kitz

  • Administrator
  • Senior Kitizen
  • *
  • Posts: 32108
  • Trinity: Most guys do.
    • http://www.kitz.co.uk
Re: VDSL modemís protocol stack - ATM?
« Reply #10 on: September 16, 2016, 10:53:19 PM »

I think we've had this discussion before, but how can there possibly be any crux point?

ATM = 5 bytes overhead (ATM cell header) + 48 bytes data
PTM = 1 byte overhead (sync octet) + 64 bytes data

It's the ATM that is largely responsible for the ADSL IP Profile being 88.2%, and the PTM for the VDSL2 IP Profile being about 96%.

To avoid serialisation delay especially with time critical applications. 
The 'crux point' being the speed at which  less overhead v increased latency caused by serialisation payoff starts to become equivalent.

Reference Ethernet in the First Mile whitepaper - link


Quote
The  latency  problem  results  from  the  fact  that  an  entire 
packet   must   be   received   for   its   content   to   become   
available  to  an  application.  In  Ethernet  terms,  ďentire 
packetĒ    means  up  to  1500  bytes,  while  in  ATM  the  unit 
of  transferred  data  is  only  53  Bytes.  On  a  typical  ADSL 
link, running at 1 Mbps, the time to transfer 1500 bytes is
12  ms.
  It  is  clear  that  queuing  alone  could  consume  a 
large   portion   of   the   delay   budget   of   a   time-critical   
application such as a voice connection. 
5. Conclusions
The  ATM-TC  and  PTM-TC  layers  defined  for  Digital 
Subscriber  Line  systems  both  have  their  advantages  and 
disadvantages  for  the  transport  of  Ethernet  frames  on  the 
local loop. At low bitrates, latency becomes an important
issue,  which  is  best  dealt  with  by  ATM,  because  of  its 
small  PDU  size.  When  different  types  of  service  are 
converged on the same channel, ATM is best equipped to
guarantee   the   QoS   requirements   of   each   individual   
service.
At  high  bitrates  latency  becomes  somewhat  less 
important, and the reduced overhead of PTM is a distinct
advantage, provided that the necessary OAM&P tools are
in  place.  The  functionality  of  the  ATM-TC  and  PTM-TC 
layers  is  comparable  in  terms  of  mechanisms  for  frame 
delineation, error detection and rate decoupling. The flow
control mechanisms are also similar, but due to the larger
granularity  of  Ethernet  frames,  larger  buffers  will  be 
required than in the case of ATM.

Logged
Please do not PM me with queries for broadband help as I may not be able to respond.
-----
How to get your router line stats :: ADSL Exchange Checker

kitz

  • Administrator
  • Senior Kitizen
  • *
  • Posts: 32108
  • Trinity: Most guys do.
    • http://www.kitz.co.uk
Re: VDSL modemís protocol stack - ATM?
« Reply #11 on: September 16, 2016, 11:49:18 PM »

I'll try explain it from a different angle this time.  Rather than trying to compare ATM v PTM previously... if this time we concentrate on just about ATM and specifically why it uses small packet sizes (despite the increased overheads) it may become clearer.

The main advantage of ATM is its small packet size.  The 53 byte cell size was designed to reduce latency/jitter over slower speeds (the adsl portion) of a point to point link.

If you have a total link whereby you start of with ethernet speeds and then suddenly hit part of the link whereby speeds are limited to say 512kbps (ie the limitation of your DSL connection), then you start getting increased delay and jitter with 1500 byte packets.

I think I gave some examples in the other post to show the effects of serialisation delay and how much time would be added to that portion of the link, but if instead this time we just look at the benefits of ATM and why it was designed to have such a small cell size.

From wikipedia

Quote
At the time of the design of ATM, 155 Mbit/s Synchronous Digital Hierarchy (SDH) with 135 Mbit/s payload was considered a fast optical network link, and many plesiochronous digital hierarchy (PDH) links in the digital network were considerably slower, ranging from 1.544 to 45 Mbit/s in the USA, and 2 to 34 Mbit/s in Europe.

At this rate, a typical full-length 1500 byte (12000-bit) data packet would take 77.42 Ķs to transmit. In a lower-speed link, such as a 1.544 Mbit/s T1 line, a 1500 byte packet would take up to 7.8 milliseconds.

A queuing delay induced by several such data packets might exceed the figure of 7.8 ms several times over, in addition to any packet generation delay in the shorter speech packet. This was clearly unacceptable for speech traffic, which needs to have low jitter in the data stream being fed into the codec if it is to produce good-quality sound.
A packet voice system can produce this low jitter in a number of ways:

    Have a playback buffer between the network and the codec, one large enough to tide the codec over almost all the jitter in the data. This allows smoothing out the jitter, but the delay introduced by passage through the buffer would require echo cancellers even in local networks; this was considered too expensive at the time. Also, it would have increased the delay across the channel, and conversation is difficult over high-delay channels.
    Build a system that can inherently provide low jitter (and minimal overall delay) to traffic that needs it.
    Operate on a 1:1 user basis (i.e., a dedicated pipe).

The design of ATM aimed for a low-jitter network interface. However, "cells" were introduced into the design to provide short queuing delays while continuing to support datagram traffic. ATM broke up all packets, data, and voice streams into 48-byte chunks, adding a 5-byte routing header to each one so that they could be reassembled later. The choice of 48 bytes was political rather than technical.[6] When the CCITT (now ITU-T) was standardizing ATM, parties from the United States wanted a 64-byte payload because this was felt to be a good compromise in larger payloads optimized for data transmission and shorter payloads optimized for real-time applications like voice; parties from Europe wanted 32-byte payloads because the small size (and therefore short transmission times) simplify voice applications with respect to echo cancellation.


So now we know why ATM was used on the slower portions of the link.   
With VDSL it doesnt matter because serialisation delay on say a 40Mbps link is negligible when you add in the benefits of the reduced overheads.

Somewhere around 12Mbps is when you start to see the benefits of PTM.   Anything less than 12Mbps is said to perform more efficiently using ATM.
Logged
Please do not PM me with queries for broadband help as I may not be able to respond.
-----
How to get your router line stats :: ADSL Exchange Checker

kitz

  • Administrator
  • Senior Kitizen
  • *
  • Posts: 32108
  • Trinity: Most guys do.
    • http://www.kitz.co.uk
Re: VDSL modemís protocol stack - ATM?
« Reply #12 on: September 17, 2016, 02:24:07 AM »

Here you go... a really nice simple explanation of serialisation delay (found by searching searialization delay) which explains everything Ive been trying to say for both this thread and the previous one. - link.

Quote
Serialization delay is the time it takes for a unit of data, such as a packet, to be serialized for transmission on a narrow (e.g. serial) channel such as a cable. Serialization delay is dependent on size, which means that longer packets experience longer delays over a given network path. Serialization delay is also dependent on channel capacity ("bandwidth"), which means that for equal-size packets, the faster the link, the lower the serialization delay.

Underneath is a simple table that shows the effect of passing 1500 byte packets through slower parts of a network.  These times are in addition to latency caused by distance etc. 

By using ATM cells with smaller packet sizes over AAL5, it gives ADSL subsubstantially reduced latency and jitter for that part of the network.
 
Think back to dial up days when it would take ~100ms to reach the 1st hop.  Distance to the ISP may still be the same, yet increased latency was caused by  serialisation over the slower part of the link.  In the early days of broadband, DSL latency was usually far better than that over a cable connection eg 20ms v 40ms because DSL used ATM.   

Going back to the other thread, what I was having difficulty explaining is that I had observed that even on VDSL, Openreach appeared to be using quite small packet sizes in the framing parameters on slower VDSL links of less than 10Mbps.   These packet sizes appeared to be similar to those you would see on ADSL stats.   As I said at that time, I wasnt quite sure where I was going with the observation other than wonder if it was in some way deliberate, in an attempt to reduce serialisation for slower lines.

I left it at "I'm not doing a good job of explaining where I was going with the rest in respect of AAL5, or rather HDLC for PTM.   Its not going to provide a solution for the op, just that there is obviously some reason why the low speed links have smaller frames."  because I dont know much about HDLC and if it can do some sort of dynamic adjustment to make lines of below a certain speed use deliberately smaller packets in an attempt to reduce the effect of serialisation.   I'm still not in a position to be able to sit and read any new and unfamiliar technical stuff... but if someone else wants to do so and follow up on HDLC then be my guest.
Logged
Please do not PM me with queries for broadband help as I may not be able to respond.
-----
How to get your router line stats :: ADSL Exchange Checker

ejs

  • Kitizen
  • ****
  • Posts: 1971
Re: VDSL modemís protocol stack - ATM?
« Reply #13 on: September 17, 2016, 07:40:59 AM »

Sigh. Yes I do understand that, but as far as I can see, it's all missing the point.

Reference Ethernet in the First Mile whitepaper - link

This whitepaper is so old it only talks about PTM using HDLC, and not this "64/65 octet" encapsulation. It's talking about VDSL1, which could have used PTM using HDLC, VDSL2 did not exist at that time.

Quote
DSL modems most often rely on ATM as
their data link protocol, a choice that was inspired by the
vision of DSL as a multi-service multi-media platform
.
ATM is indeed very well equipped to handle multiple
streams of data, each with different bandwidth and quality
of service (QoS) requirements.

But we never used DSL like that. There aren't different classes of ATM cells carrying voice data and Internet data. Everything sent over the ATM used over an ADSL link goes into the same PPP / Internet data traffic, there's no QoS at the ATM level.

If you are sending a 1500 byte packet over ATM over AAL5, it'll still take the same amount of time to transfer all those 1500 bytes due to the bandwidth of the link! If you are sending a 100 byte packet over ATM/AAL5, it'll take three ATM cells then the end of the AAL5 trailer. Sending the 100 byte packet over PTM, it'll take two of those 64 byte PTM codewords, and the sync byte will indicate it's the last codeword of the packet.

Quote
ATM broke up all packets, data, and voice streams into 48-byte chunks

But going over the ATM used over an ADSL link, all there is is data packets. A lot of the information about ATM in general, is, as far as I can see, not relevant for how it's used on ADSL. When it's talking about low delays for voice traffic, that'll be from putting the voice traffic in separately classed cells from the bulky data traffic.

And the PTM used on VDSL2 is breaking up everything into 64 byte chunks anyway!

When I'm referring to this 64/65 byte encapsulation, that's how it's written in the documentation, it's not that I don't know whether it's 64 or 65 bytes. Most of the documentation about general ATM pre-dates this 64/65 byte encapsulation used on VDSL2.
Logged

ejs

  • Kitizen
  • ****
  • Posts: 1971
Re: VDSL modemís protocol stack - ATM?
« Reply #14 on: September 17, 2016, 07:44:25 AM »

Is this relevant from a cut & paste ?
...
5. PPPoA has slightly less overhead than PPPoE therefore, it is a bit faster.

No, it's not relevant. Because when you use PPPoE over ADSL, it's still being put into ATM cells over the ADSL link, so the PPPoE is really PPPoEoA, which is only adding extra overheads compared to having PPPoA directly. Unless you were using PTM over ADSL2, which nobody does.
Logged
Pages: [1] 2