WO2005034477A1

WO2005034477A1 - Communicating method using scheduling of transmissions

Info

Publication number: WO2005034477A1
Application number: PCT/GB2004/004147
Authority: WO
Inventors: Kevin HOBBIS
Original assignee: Hutchison Whampoa Three G Ip (Bahamas) Limited
Priority date: 2003-10-03
Filing date: 2004-09-29
Publication date: 2005-04-14
Also published as: GB2406743A; GB0323183D0

Abstract

A method of scheduling communication transmissions in telecommunications networks. Functionality is provided at the application level which predicts when time critical communications will occur. On the basis of such predictions, other communications are sent at times when they will not interfere with time optimal communications. In particular, the method is a method of controlling the transmission of communications from user equipment in a communications network in which there are a first class of communications which may be sent from user equipment which are time critical and which must be sent and/or will elicit responses from remote equipment at predictive times, and a second class of communications the sending of which from user equipment is not time critical; the method comprising: determining time periods during which communicates in said first class or said responses are likely to occur; and scheduling transmission of communications in said second class such that they do not occur during said determined time periods.

Description

COMMUNICATING METHOD USING SCHEDULING OF TRANSMISSIONS The present invention relates to telecommunication networks and in particular to the scheduling of communication transmissions in such systems. In a telecommunication system, such as a mobile telephone system, there are provided, as is well known, a network infrastructure, which may include wired and wireless communication links, and user devices which connect to the infrastructure in order to communicate with each other. Recent generations of mobile telephones have seen handsets develop to a position in which they are provided with much more functionality than a traditional telephone and are, in some ways, more akin to personal computers on which a number of different program applications may operate. The present invention has been developed in the context of such systems, but it will be appreciated that it is, for the reason just mentioned, also applicable more generally to computer systems and networks. In devices of the type mentioned above it is further well known for a plurality of different applications to be operating or running on a user device simultaneously. These applications may be user applications of which the user is aware but also may include background or housekeeping applications which the device runs. Some or all of such applications may at various times wish to send communications which would typically be addressed to corresponding applications at other user devices or at server devices in the network. Communications to and from the network are directly handled by specific functionality within the device (typically referred to as communication layers) which may handle the proper addressing of the communication and processing of the data to ensure compliance with specific communication protocols etc which are in operation. It is known within such layers to provide simple scheduling functionality to enable proper handling of multiple communications received from the applications at the same time. This may include buffering and prioritisation based on well known schemes. Such handling of communications can present problems, as will be explained on the basis of an example in the context of a mobile telecommunication system. In such a situation a user device (the mobile handset) may have one radio link or bearer via which to communicate with the network and thereby other users and services. The handset may however have a number of different sessions set up, such as a voice call and an instant messaging service. Communications relevant to all of the established settings are sent using the same radio bearer and therefore scheduling of the various messages onto the bearer is required. As mentioned above simple scheduling may be performed by the communication layers to ensure that only one message at a time is sent to the radio bearer. However problems may still arise. In such systems, the protocols for establishing the session may require that, in order to continue a session, the user device must re-register at certain intervals. This is performed as a housekeeping function in the background as mentioned above. If the device is unable to send or receive such messages, sessions may become disconnected causing inconvenience to the user. It has been found that such a problem can arise in the circumstance where an application has to generate communications on the radio bearer at critical times relating to such registration. For instance, an Instant Messaging application may be sending a message with a large attachment which may occupy the radio bearer for a considerable period of time. If during that period, a time comes when another application needs to register or re-register a session that session may be lost. Normal scheduling procedures in the communication layers cannot overcome such problems. The present invention provides functionality at the application level which functions to predict when time critical communications will occur. On the basis of such predictions, other communications are sent at times when they will not interfere with the time optical communications. In particular the invention provides a method of controlling the transmission of communications from user equipment in a communications network in which there are a first class of communications which may be sent from user equipment which are time critical and which must be sent and or will elicit responses from remote equipment at predictable times, and a second class of communications the sending of which from user equipment is not time critical; the method comprising: determining time periods during which communications in said first class or said responses are likely to occur; and scheduling transmission of communications in said second class such that they do not occur during said determined time periods. The present invention will be better understood from the following description of a preferred operational embodiment in the context of SIP operation The description is given in conjunction with the accompanying figures, in which: Figure 1 illustrates simplified session set up using SIP; and Figure 2 illustrates scheduling an accordance with this invention implemented by SIP. The introduction of the Session Initiation Protocol (SIP) into mobile networks for multimedia session control, Instant Messaging, Presence and Conferencing greatly enhances the customer experience. However, in networks where the access speed is restricted e.g. mobile networks where the air interface resource is at a premium, there is a potential foi quality of service issues to arise (e.g. call set up delay) due to the sharing of the same bearer for the SIP call control messages and possible IM/Presence messages. SIP call control uses a specific set of messages and responses, defined in RFC3261, to allow a user to set up sessions (for voice, video and other usage). Extensions to SIP signalling are defined to allow for Instant Messaging and Presence support using the same SIP signalling network. The characteristics of Instant Messaging and Presence (performance and message size) are not necessarily compatible with the requirements for call control. An Instant Message may contain a large attachment (image or video clip for instance), which, on a restricted bandwidth link may take a significant amount of time to transport. During this time it would clearly delay any further messages, and for instance a session set up message could easily be delayed by such a transfer, resulting in the user aborting the session, or dissatisfaction due to extended set up times. SIP signalling uses a well defined sequence of requests and responses, see figure 1, for a simplified overview.

The steps are as follows 1. User Agent 1 sends an Invite (with offered session parameters). This is passed by the SIP Proxy to User Agent 2 using SIP routing rules etc. 2. The 100 trying messages are used to keep the originator informed of session progress. 3. The 183 Session Progress contains an answer to the offered session parameters. 4. The PRACK contains the proposed agreed session parameters. 5. The 200 OK indicates a successful session set up.

These requests and responses form well known bounded procedures. There is one other key factor about SIP signalling, in that some Requests can contain an Expires header which indicates, depending on the request type, how long that request is valid for.

Examples are as follows 1. REGISTER : in a register messages the expires time indicates the time after which the user needs to re-register 2. SUBSCRIBE : this indicates how long a subscription (e.g. to another users presence information) is valid for, and therefore when it needs to be refreshed. 3. PUBLISH : this indicates the minimum time before the next publication of updated presence information will take place (if no change has taken place then no publication will be needed) In all cases the Expires time is proposed by the terminal and agreed/set by the network. The operation of this embodiment uses the fact that the SIP procedures are well bounded and that many functions have an agreed refresh time to allow SIP signalling to be scheduled at the application (i.e. SIP) layer. Figure 2 illustrates some examples. The principle applied is that the User Agent knows about the times agreed via the expires header for different procedures, and therefore at what point it will have to perform more signalling to refresh these. It can therefore set a guard period around this end time when it will not initiate new signalling, and in effect it can anticipate if any proposed signalling, if started, will encroach into the guard period. It is possible to initiate new signalling at any time outside of these guard periods. A similar stance can be taken for the period between an initial request and the 200 OK response (or other final response e.g. in the case of unsuccessful or rejected requests), i.e. no new signalling is initiated during an ongoing procedure, so a guard period is also set between the request and the final response. SIP INVITEs will be initiated due to user interaction, and cannot be predicted, r so it is possible that they may be requested during a guard period. In the case of a user initiated INVITE request the guard period can be overridden - to maintain user satisfaction. The basic principle is illustrated Figure 2, with the request to send an Instant Message, as described below. 1. The size of the instant message (including attachments) will be known, so it is possible for the terminal to estimate the transmission time. 2. On receipt of the request the terminal identifies that there is a guard period already running due to the ongoing Publish procedure, hence the sending of the IM is deferred. 3. On completion of the Publish procedure the IM is still pending. Again the terminal determines that transmission would encroach upon the upcoming guard period identified from TI fro the necessary re-registration. 4. The re-registration procedure takes place and although the guard period for TI runs out, a new guard period for the second part of the authenticated registration procedure is still running, so the IM is still not sent. The register has an authorisation request from the network requiring a second register procedure. This process illustrates why the guard period may be set during the Register procedure. 5. Upon completion of the registration the terminal again looks at the pending IM request. In this case the transmission time will not overlap with any already identified guard periods, so the DV1 can be scheduled and sent.

Guard periods are set between the initial and final response for all procedures so that for instance the transmission of the IM did not take place between the REGISTER (or more importantly an INVITE) and it's 200OK because it may delay messages transferred before the 200OK as part of these procedures (e.g. an authenticated register as shown in the illustration, or signalling due to unsuccessful procedure. Or detailed INVITE signalling illustrated in a simplified way in Figure 1). The example shown is relatively simple, and for instance the setting of the guard periods can be more sophisticated e.g. to include other constraints on the refresh procedures e.g. the refresh may be required to start some minimum fixed time before expiry. The application/session layer will be aware of these restrictions and can account for them when setting the guard periods.

Claims

1. A method of controlling the transmission of communications from user equipment in a communications network in which there are a first class of communications which may be sent from user equipment which are time critical and which must be sent and/or will elicit responses from remote equipment at predictable times, and a second class of communications the sending of which from user equipment is not time critical; the method comprising: determining time periods during which communications in said first class or said responses are likely to occur; and scheduling transmission of communications in said second class such that they do not occur during said determined time periods.

2. A method according to claim 1 in which said first class of communications includes Session Initiation Protocol (SIP) invite and registration communications.

3. A method according to claim 1 or 2 in which said second class of communications includes Instant Messages (IMs).

4. A method according to any of claims 1-3 in which said time periods are determined based on known refresh times of registered procedures.

5. A method according to any of claims 1-4 in which said time periods are determined to extend by predetermined amounts before and after actual expected times of communications.

6. A method according to any of claims 1-5 in which said scheduling of communications in said second class is done taking account of expected transmission times of said communications.

7. A method according to any preceding claim in which the expected time of a future communication in said first class is predicted on the basis of information contained in an earlier communication in the first class.