November 8, 2012

Why Fax over IP Fails

Unified communications (UC) often falls short of complete unification when the customer keeps a plain old telephone service (POTS) line for each facsimile machine.  Makes sense because fax over IP isn't as reliable as over the public switched telephone network (PSTN).  Fax almost always works on the legacy PSTN because the modems in fax machines were optimized by design for an analog channel.  If the path includes an IP segment, many things can go wrong.  Here's why.
 
The procedures (defined in ITU T.30) compensated for well-known impairments in voice lines:  noise, dropouts, fading.  Error checking and forward error correction covered short spikes of errors.  Retransmission and dropping down to a lower modem speed overcame persistent noise and distortion.

How to Minimize IP Impairments

IP packets create an environment far different from that of the original fax machines.  The most sensitive impairment for IP is packet loss, something the modems can't overcome very well.  Depending on the method chosen for a fax call, a packet loss ratio of more than 1% could cause the call to drop.  Fortunately that quality is within reach if the design for an IP network addresses packet loss aggressively. 

Current best practices to minimize loss aim at avoiding congestion:
  • Ensure adequate bandwidth on all link segments that will carry voice or fax.
  • Prioritize fax and voice with class of service configurations in LAN switches. 
  • Obtain priority service classes from WAN carriers with Service Level Agreements.
  • Select queuing algorithms that always put voice and fax first.
How much bandwidth a fax connection requires depends on how it is formatted on the IP network.

How to Avoid Synchronization Problem

The second change from analog or TDM transmission to packet transmission is the loss of synchronization.  Modems play a sound continuously so the receiver knows how fast the sender transmits.  Packets interrupt the bit stream, losing continuity.  The receiver can base its timing on a local free-running clock which differs slightly from the sender's clock rate.  If the fax is long enough, the receive buffer underflows or overflows, breaking the connection. 

To handle faxes with a modest number of pages you can set a largish and fixed-size jitter buffer at the receiving end to absorb differences in clock frequencies between send and receive gateways.   Dynamic management of the jitter buffer by varying the clock rate avoids this problem.

The International Interconnection Forum For Services Over IP (i3 FORUM, www.i3forum.org) has researched fax over IP for several years.  Its latest report on the topic (Technical Specification for Voice over IPX Service, Release 3.0, May 2012) [Link to http://i3forum.org/wp-content/uploads/2012/05/i3F-Technical-VoIPX-Release-3-FINAL-2012-5-3.pdf] spells out the bandwidth requirements for various codecs, packet accumulation times, and transmission methods.  The report also offers guidelines for maximum latency and other parameters. 

i3Forum identified another cause for the failure of T.38 fax calls in the way fax gateways process certain session initiation protocol (SIP) messages.  Specifically, a T.38 gateway at the receiving fax machine will send a REINVITE to the sending end when it realizes the call is for fax.  The REINVITE requests a change from the voice codec--usually pulse-code modulation (PCM)--to T.38.  If the sending gateway hears the answer tone from the called fax machine (on the initial PCM voice connection) before the REINVITE arrives, it will drop the PCM connection and attempt to set up a new IP connection for fax.  This move drops the voice call which looks like a call failure to fax machines so they hang up.

One of the participants in the I3F testing has applied for a patent on one solution.  Commetrex Corporation's solution allows the sending fax gateway to refuse the REINVITE if it arrives after the answer tone.  The call stays with the PCM/G.711 codec and uses "fax pass-through" for the entire transmission.

As solutions work their way into fax/SIP protocol stacks in the not distant future we should see much more reliable fax over IP.

[T.38 is covered in Mr. Flanagan's latest book, VoIP and Unified Communications (Wiley, 2012).]


4 Comments

Do we really need to continue to worry about fax in a UC environment? It’s an antiquated format – almost as old as telex. Why not use UC as a convenient excuse to stop using a format that is easily replaced by digital file attachments?

Telex service is still in use because it and fax offer certain features that some customers want. For example, Telex verifies the users address number; fax conveys signatures with authority. Both could linger for years more.

You mention a proposed solution by Commetrex. Where does this solution “fit” in the network? Is it software that goes in the SBC? Will it be a separate device of some form?

Commetrex software could reside in an SBC, analog terminal adapter, router, server, or any fax termination equipment.

Search Webtorials

Get E-News and Notices via Email


  

 



  

I accept Webtorials' Terms and Conditions.

Trending Discussions

See more discussions...

Featured Sponsor Microsites






















Archives

Notices

Please note: By downloading this information, you acknowledge that the sponsor(s) of this information may contact you, providing that they give you the option of opting out of further communications from them concerning this information.  Also, by your downloading this information, you agree that the information is for your personal use only and that this information may not be retransmitted to others or reposted on another web site.  Continuing past this point indicates your acceptance of our terms of use as specified at Terms of Use.

Webtorial® is a registered servicemark of Distributed Networking Associates. The Webtorial logo is a servicemark of Distributed Networking Associates. Copyright 1999-2018, Distributed Networking Associates, Inc.