TELECOM ACCESS STANDARDS NEWSLETTER NO. 136 NOVEMBER 2002 CONTENTS 1. TELECOM NETWORK SUPPORT FOR V.92 OPERATION 2. WHAT DOES ALL THIS MEAN TO A CUSTOMER? 3. POTENTIAL CPE COMPATIBILITY PROBLEMS 4. TESTING PROCESS 5. ADDRESSING THE POTENTIAL PROBLEMS 6. UPGRADING EXISTING V.90 MODEMS 7. INDUSTRY COMMENT 8. APPLICATION OF INTERIM PTC SPECIFICATIONS RETURN TO MAIN INDEX 1. TELECOM NETWORK SUPPORT FOR V.92 OPERATION Telecom has been working for some time on the introduction of V.92 modem support on its Network Access Servers (NAS) so that it can offer the various new features of V.92 to ISP's. The network side of the testing is almost completed and this service is expected to go ahead very shortly. Newsletters 131 and 132 and, in particular, to the Appendix to Newsletter No 132, supplemented the general requirements of PTC 200 and covered initial testing proposals for V.92 modems. This draft was based on the situation that our network did not then support the service, so testing had to be somewhat restricted. We are now developing a revised schedule of tests that will confirm the operability of the new V.92/V.44 features, which may include the following:- * Quick connect (by "remembering" the modem settings previously negotiated for the line concerned and assuming there is no need to change these settings for a subsequent internet session); * PCM "upstream" (providing for up to 48 kbit/s upstream, although this speed is not expected to be achieved on the initial release of our network equipment software); * V.44 compression (around 25% faster file transfer claimed, with even more effective speed increase for some classes of file); * "Suspend/Resume" (which, with a facility for detecting "Call Waiting tones, provides for a Modem on Hold (MOH) feature and the possibility of suspending an internet session for some specified period to accept an incoming telephone call, then resume the session when that call has been completed). Many prospective modem purchasers will be aware of these four features, but it must be stressed that:- * All are not necessarily supported by the network or the ISP concerned. Even when both entities support the features, they may not be operable in all circumstances. * To avoid confusion and customer complaints that some expected features do not work, both suppliers and ISP's need to accurately define exactly what features their products and services support. * In accordance with the Consumer Guarantees Act, clear warnings should be given by modem suppliers where there is any doubt about the effectiveness of the features claimed for a particular product. The first three features mentioned above are modem/NAS compliance issues and we will need to define some additional PTC test requirements to make sure these features work as planned by any modems being submitted for Telepermit. Telecom is carrying out some initial compatibility testing and we expect to use the results to further develop the interim supplement to PTC 200, which applies to all products claiming V.92 compliance. 2. WHAT DOES ALL THIS MEAN TO A CUSTOMER? At present, most internet users have a 56 kbit/s modem and a single line. The "heavier "users, who send a lot of time "on-line" will frequently take a second line to handle their phone calls, especially if they have a family also wanting to make calls. Those in the next level up will generally go for "JetStream" as this leaves the phone line available and provides much faster internet service at little more cost than a second line. V.92, in conjunction with Telecom's "Call Waiting" Smartphone service, offers a further option - as long as the modem and ISP support the "suspend/resume feature. While the phone is not available for outgoing calls during an internet session with a suitably set up V.92 modem, incoming calls will result in a Call Waiting signal which can be detected by the modem. This results in a screen message to indicate to the user that "An incoming call is waiting - Do you wish to accept it? YES or NO" If the user clicks on "YES", the modem signals the network equipment to negotiate suspending the internet call, then causes a switchhook flash (a brief line break) to swap over to the phone call. If all goes well and the phone conversation is completed before the suspend limit is exceeded, the user can resume the internet call. This provides a low cost means of being available to answer incoming calls and still have more or less continuous access to the internet. However, there may be problems as described below. The other big advantage of V.92 is "Fast connect". Claimed to be about twice as fast as a conventional V.90 modem, even if the internet call has to be curtailed because of a long conversation, the modem can reconnect to the internet a lot quicker. "PCM Upstream" will be useful to those who tend to spend more than the average amount of time uploading files to the internet. Initially, the speed is not likely to achieve much more than the usual best speed of a V.90 modem at 33.6 kbit/s. However, not all V.90 users can get this speed and there will be circumstances where this new feature makes a significant difference in up-loading performance. The effectiveness of the V.44 compression feature depends on the nature of the files being downloaded. V.44 is claimed to be the "next generation of file compression" and will thus replace the V.42bis compression used by most V.90 modems. HTML files are expected to show the best file transfer speed gains with V.44. There is quite a lot of background information on V.92 operation available from the internet. A useful reference website is www.v92.com However, when reading this information it is important to note that not all the features and facilities it covers are available here. One example of this is "Caller Display with Call Waiting". Telecom does not provide this service and we will ask modem suppliers to make this quite clear in their marketing brochures and user manuals. 3. POTENTIAL CPE COMPATIBILITY PROBLEMS Based on initial considerations of telephone and modem dc current and voltage characteristics, we are concerned that the MOH feature could result in operating and compatibility problems for many users. The average modem has a very low voltage across it when on-line, typically around 5 - 6 V at 20 mA. This is deliberately designed to minimise heat dissipation. On the other hand, many modern electronic phones need more than 6 V across them and around 20 mA available for them to "turn on'" and operate correctly. This is not usually a problem, as we do not generally need to connect two items on line at the same time, other than for brief periods of handover between two telephones and, until V.92 introduction, it would be very unusual to have a phone and a modem connected at the same time. In the V.92 MOH mode, both modem and phone may be across the line together for at least a short period. Where the line current is around 20 - 30 mA due to long lines or lines with restricted power feeds, there could be current sharing problems and insufficient voltage or current to operate the telephone correctly. To make matters worse, the modem shunts the phone with roughly the same ac impedance, probably dropping voice levels to/from line by about 3 dB and increasing sidetone at the same time. Higher sidetone usually causes the user to drop his/her voice as well, adding to the degraded voice performance for the other party to the call. It appears certain that there will be at least some combinations of CPE used by some customers that either won't work or suffer seriously degraded voice quality if both CPE items are across the line at the same time. There are hundreds of different phone types and thousands of modem types now Telepermitted, so the number of possible combinations is vast. The risk of at least some suspect combinations is thus high. With the large number of types of products in service, it is impossible to check out every possible combination that could arise in practice. An added complication is the validity of the operating sequence, as was explained in the above newsletters. Whether or not detecting the Telecom version of the call waiting signal, holding an internet call and transferring to and from the voice call works correctly is another area of doubt that needs to be checked by a test laboratory. On the more positive side, the MOH feature arises from the "suspend - resume" capability in V.92, which is made even more user-friendly by the fast train feature. There is no specific guidance in ITU Rec. V.92 as to how Call Waiting/MOH should be implemented, other than for the very low level commands the modem sends to the NAS to suspend and then resume a session. The problems with double termination do not occur inherently. If the modem contains suitable hardware, it could switch between itself and the phone so that only one device is terminating the line at any time. Unfortunately, it is likely that most V.92 modems will be simply new firmware in existing V.90 hardware, so the voice performance is likely to be compromised to at least some extent. Telecom's own test staff are looking into these issues and, if necessary, we will require modem suppliers to include a warning to customers that the service may not work with some combinations of phone and modem. 4. TESTING PROCESS Local Telepermit testing laboratories have been asked to carry out practical testing to the draft PTC 200 supplement of Newsletter No. 132 on all V.92 modems received for Telepermit evaluation. Each supplier of a V.92 capable modem is required to give clear instructions to users on how the MOH sequence is intended to work. The lcal test lab can then confirm whether or not this sequence does work in the Telecom network environment. Overseas test labs can go part of the way in V.92 modem testing, but their network conditions are unlikely to match ours exactly. This means V.92 modems may have to be covered by a mandatory local test to supplement overseas PTC 200 test reports. 5. ADDRESSING THE POTENTIAL PROBLEMS Where there is any doubt about the capability of a V.92 modem to perform any of the new features of the protocol, it can be argued that the Consumer Guarantees Act requires the Telepermit holder to include clear warnings to this effect. In line with this, and mainly to minimise the risk of customers complaining to Telecom if something does not work as expected, we will require warnings to be included in V.92 user manuals. With respect to the MOH issue, the following would appear suitable:- "Due to the differing characteristics of the many modems and telephones currently in service, no assurance can be given that the MOH feature will work correctly. Telephone performance may be significantly degraded in this mode. Do NOT call Telecom if this feature fails to operate correctly with your telephone". Needless to say, such a warning is not applicable if testing reveals that the proposed operating sequence does not work at all on the Telecom network. In such cases, the Telepermit holder will be required to include a clear warning to the effect that:- "Modem on Hold is not available with this product". Consideration is also being given to the inclusion of a recommendation in the user manual, as follows:- "In those cases where the MOH operating sequence does work and the telephone connects, but speech quality is significantly degraded, it is recommended that this mode be used only for brief calls. Where a long conversation is likely, it is recommended that the modem be forced to disconnect. This will improve the quality of the voice call". These issues will be dealt with in a revised version of the interim PTC 200 supplement originally published with Newsletter No. 132. In the meantime, this interim supplement is to apply for V.92 products and their related testing. 6. UPGRADING EXISTING V.90 MODEMS Access Standards requires that holders of Telepermits for V.90 modems still being offered for sale and upgraded to V.92, are to have the operation of their products checked out by a test laboratory in line with the interim PTC 200 supplement. This should be done before such products are offered for sale in "V.92 mode". Appropriate warnings and recommendations are to be given with the upgraded product, as for any other new V.92 modem. There is a potential problem where existing privately-owned V.90 modems can be up-graded on-line by the user without Telecom or the Telepermit holder being aware of the change. Where such modifications cause service problems, the only practicable approach is that the "customer is on his own". In such circumstances, the Telepermit may be invalidated if the modem no longer works correctly on the Telecom network. Access Standards will endeavour to publicise this issue, especially if frequent customer complaints arise as a result of such "upgradings". 7. INDUSTRY COMMENT V.92 is a new development in New Zealand, although it has been in use in some overseas countries. As explained above, it presents some new issues for the Telepermit system and for suppliers of these products. Because customers may well expect any "V.92 modem" to support all the new V.92 features they read about in magazines or see on the internet, it is important that there is a clear definition of just what features are supported by each product. From a supplier's viewpoint, it is also important that there is no suggested obligation to provide all such features, when they had not intended to do so. Any comments or suggestions from readers on the testing process or our proposed means of dealing with V.92 modem issues will be most welcome. 8. APPLICATION OF INTERIM PTC SPECIFICATIONS I note that some test laboratories and suppliers have not been clear when interim PTC specification supplements published in these Newsletters come into force. To clarify the position, the following general rules cover our intentions:- 1. Where the Newsletter article is written in ordinary text and simply explains or outlines a new product, feature or service, this is essentially information for suppliers, test laboratories and customers to make them aware of the matters discussed. Such information complements or provides an interpretation of the existing PTC specifications concerned, but does not change them. 2. Where the Newsletter article publishes an interim supplement or amendment to a PTC Specification, either referring to specific clause numbers or formatted as a new set of requirements, these requirements come into force on the date the Newsletter is published unless a later date is notified in the Newsletter concerned. In general, we intend to give at least 3 months notice of existing requirements that are being changed in order to give any suppliers who are part way through their product design and testing process opportunity to incorporate those changes. Where a new service or feature is being introduced and it is not covered specifically by any existing PTC clauses, the aim is to address potential problems as quickly as possible - preferably before any problems impact on Telecom, the suppliers concerned or the customers who purchase their products. For those overseas test laboratories more used to a formal system of specifications or standards that are published by a national or international standards body, our system may seem rather casual. Nevertheless, it allows us to react quickly to new issues, as in the example of this Newsletter on V.92 introduction illustrates. DOUG BURRUS Manager Access Standards