METHOD AND DEVICE FOR PROVIDING SCALABILITY IN STREAMING/ARCHIVING SYSTEMS FOR CONFERENCE CALLS

Field of the invention

The present invention relates to videoconferencing and streaming/archiving systems.

Background of the invention

To have a meeting among participants not located in the same area, a number of technological systems are available. These systems may include videoconferencing, web conferencing or audio conferencing.

The most realistic substitute of real meetings is high-end videoconferencing systems. Conventional videoconferencing systems comprise a number of end-points communicating realtime video, audio and/or data streams over WAN, LAN and/or circuit switched networks. The end-points include one or more monitor (s), camera (s), microphone (s) and/or data capture device (s) and a codec, which encodes and decodes outgoing and incoming streams, respectively. In addition, a centralized source, known as a Multipoint Control Unit (MCU) , is needed to link the multiple end-points together. The MCU performs this linking by receiving the multimedia signals (audio, video and/or data) from end-point terminals over point-to-point connections, processing the received signals, and retransmitting the processed signals to selected end-point terminals in the conference.

By using a videoconferencing system, e.g. a PowerPoint presentation or any other PC-presentation may be presented while still being able to see and hear all the other participants .

In an end to end visual communications solution it is advantageous for video conferences to be made available to

wider audiences which may have entirely different time constraints. An example of this is an important meeting or announcement that needs to be available to all employees in a multinational company. This involves multiple time zones and many scheduling conflicts. A solution to this is a video conferencing recording and streaming system that can record video conferences for delivery at a time of the viewers choosing. For people who don't need to be directly involved in the conference the system can also stream it live as it is happening. The delivery method referred to here is streaming over the internet.

Presenting multimedia content by streaming data to computers through a web interface is well known. The data stream may be transmitted in real-time, or a play back of an archived content through a streaming and/or archiving system. Conventional streaming data is adapted for storage and distribution, and therefore the multimedia content is represented in a different format than for video conferencing. Hence, to allow for streaming and archiving of a conventional video conference, a system for converting the multimedia data is needed.

A streaming and/or archiving system for video conference calls is preferably provided with a network interface for connecting the device to a computer network, audio/video and presentation data interfaces for receiving conference content, a file conversion engine for converting presentation content into a standard image format for distribution, and a stream encoder for encoding the content into streaming format for distribution. The streaming and/or archiving system is further equipped with a stream server for transmitting the encoded audio/video content and a web server for transmitting web pages and converted presentation content to terminals located at nodes of the network, he streaming and/or archiving system is also adapted to create an archive file consisting of the encoded stream data, residing at local storage media or in a

server/database, to enable later on-demand distribution to requestors at remote terminals over the computer network.

However, streaming and/or archiving system as discussed above have limited amount of resources. Whether it is implemented in software or hardware there will be a maximum amount of simultaneous video conferences the device can handle. This number is inversely related to the complexity of the recording/streaming task. That is, the more CPU intensive the recording and streaming tasks, the less of them the system will be able to handle simultaneously. In many cases however, these kinds of CPU intensive tasks produce more desirable outputs. For example encoding normal and extended video sources into a streaming format such as Windows Media is resource intensive, but produces an output that can be delivered to a large number of people over a range of bandwidths and without the need for additional plugins or codecs to be installed on client computers.

Deployment of a videoconference distribution device in large installations may require many more simultaneous outputs than a single device can produce. Adding more systems to increase resources is a logical step, but also increases the complexity of dealing with the total solution. All devices must be managed separately, so carrying out administration tasks for each of the deployed standalone devices involves a duplication of effort on the part of the administrative users. Each standalone device has its own Graphical User Interface (GUI) which increases the probability of end users trying to access the wrong interface to perform routine tasks, such as creating and gaining access to the recorded/streamed content.

One way to deal with multiple standalone streaming and/or archiving system is to have an intermediate management system that manages the interaction between all the standalone devices, as shown in figure 1. However, this requires a separate system altogether and provides one

point of failure. In addition, the streaming and/or archiving systems are still treated as individuals by the video conference systems. Each one has its own addresses, and a user must know several different addresses to connect to in order to utilize all devices. Further, if the first streaming and/or archiving system dialled by the user is out of resources and unable to handle the call, the user must start all over and dial one of the other streaming and/or archiving systems.

Hence, current solutions involve managing all systems independently, or using another external control system to manage them all. The disadvantage to this is that the task of interacting with and administering the devices becomes more complex and time consuming as the number of devices increases.

Summary of the invention

It is an object of the present invention to provide a system and a method solving at least one of the above- mention problems in prior art.

The features defined in the independent claims enclosed characterise this system and method.

Brief description of the drawings

In order to make the invention more readily understandable, the discussion that follows will refer to the accompanying drawings,

Figure 1 is a block diagram illustrating an intermediate managing system for managing a plurality of streaming and/or archiving systems,

Figure 2 is a block diagram illustrating a typical streaming and/or archiving system according to prior art,

Figure 3 is a block diagram illustrating a typical system architecture of audio and/or video conferencing equipment in a packet switched network,

Figure 4 is a block diagram illustrating typical system architecture of a content server cluster according to the present invention,

Figure 5 is a block diagram illustrating typical system architecture of a cluster controller,

Figure 6 is a block diagram illustrating typical system architecture of a content server cluster according to one exemplary embodiment of the present invention,

Figure 7 is a block diagram illustrating typical system architecture of a content server cluster according to another exemplary embodiment of the present invention

Detailed description the invention

In the following, the present invention will be discussed by describing a preferred embodiment, and by referring to the accompanying drawings. However, people skilled in the art will realize other applications and modifications within the scope of the invention as defined in the enclosed independent claims .

The present invention is a method and system for clustering two or more video conferencing recording/streaming devices, making the cluster of two or more devices appear as one single device for the user.

Figure 2 show video conferencing equipment connected to a typical packet based network. H.323 is an International Telecommunications Union (ITU) standard that provides specification for computers, equipment, and services for multimedia communication over packet based networks that

defines how real-time audio, video and data information is transmitted. The H.323 standard specifies four kinds of components, which, when networked together, provide the point-to-point and point-to-multipoint multimedia- communication services:

1. terminals (101)

2. gateways (102)

3. gatekeepers (103)

4. multipoint control units (MCUs) (104)

H.323 Terminals are the endpoints on the LAN that provide real-time two way communications. Terminals are usually a personal computer (PC) or a stand alone-device (e.g. video conferencing endpoint) . Gatekeepers are responsible for providing address translation between an endpoints current IP address and its various H.323 ID aliases, call control and routing services to H.323 endpoints, system management and security policies. An H.323 gateway provides connectivity between an H.323 network and a non-H.323 network (e.g. ISDN). Finally, the MCUs provide support for conferences of three or more H.323 terminals. Further a video conferencing streaming and/or archiving system (106) is connected.

As shown in figure 3, a video conferencing streaming and/or archiving system (106) , hereafter referred to as a content server (CS) , is preferably provided with a network interface (204) for connecting the device to a computer network, audio/video and presentation data interfaces for receiving conference content, a file conversion engine for converting presentation content into a standard image format for distribution, and a stream encoder for encoding the content into streaming format for distribution. The CS is further equipped with a stream server for transmitting (203) the encoded audio/video content and a web server for

transmitting web pages and converted presentation content to terminals located at nodes of the network. The content server is also adapted to create archive files (202) consisting of the encoded stream data, residing at local storage media or in a server/database, to enable later on- demand distribution to requestors at remote terminals over the computer network.

According to a typical mode of operation, the conference is initiated by including the content server (106) as a participant in the conference. The content server (106) accepts or places H.323 video calls as point-to-point (only one H.323 system in the call, typically used to record training materials from a single instructor) or multipoint (2-n H.323 systems in the call, typically used to stream or archive meetings) . Viewers at remote terminals (PC_1 -

PC_n) can access a conference by directing a conventional web browser to an URL (Uniform Resource Locator) associated with the content server (106) . After completion of validation data interchanges between the viewer and the distribution device, the viewer is able to view the personal interchange, i.e. the conversation and associated behaviour, occurring between the participants at the conference presenter site, as well as view the presentation content being presented at the conference site. The multimedia content is viewed in a multiple-window user interface through the viewer web browser, with the audio/video content presented by a streaming media player, and the presentation content displayed in a separate window. When requested by the head of the conference or by the conference management system, encoded stream data is stored in a server as an identifiable file (202) .

As mentioned above, in order to produce streaming and recording outputs from a video conference, the content server (106) acts like a video conferencing endpoint (terminal) . The CS (106) can join a point-to-point or a multipoint video conference just like any other endpoint,

but instead of displaying the output on a screen like most regular video conferencing endpoints, it processes the data into other useful outputs.

As an endpoint the CS may also have one or more alias addresses associated with it. The alias addresses provide an alternate method of addressing the endpoint. These addresses include E.164 (network access number, telephone number, etc.), H.323 IDs (alphanumeric strings representing names, e-mail-like addresses, etc.), and any others defined in Recommendation H.225.0. Alias addresses are unique within a Zone, Domain, among Zones, and among Domains.

As discussed in the background section above, a CS is limited in the number of outputs it can create. To increase the number of outputs, additional devices must be added. Increasing the number of content servers consequently increases the complexity of managing and interacting with them. However, according to the present invention the individual standalone content servers are clustered together in such a way that the user experience is as if they were interacting with a single system, no matter how many systems are part of the cluster. In addition to this, administrative tasks can be performed in the same way for the cluster as for a single distribution device.

Figure 4 is a schematic drawing of a content server cluster according to the present invention. Two or more content servers are joined to form a cluster, and in a cluster with m content servers (106), one of the CS' s assumes the role of the cluster controller (401) , while the remaining n=m-l devices assumes the role of nodes (101/1 - 101/n) . There are two main configuration files for the content servers, one for the web based user interface and one for the call handling unit. The cluster settings reside in the user interface's configuration file and describe all the machines in a cluster, e.g. whether they are a node or a

controller, and for the controller whether it can record and/or archive as well as just be the user interface. All the content servers in a cluster are initially identical. By changing the configuration file on a content server assigning it to be a node, disables the user interface on this content server. Furhter, changing the configuration file on a content server, assigning it to be the a controller, configures it to propogate its system settings to all the nodes in the cluster. Consequently, the cluster controller (401) becomes the user's only point of interaction with the cluster, meaning that even though there are several devices in the cluster, only one interface (for device settings, recording and/or archiving and viewing streams/recordings, etc.) is provided. Further, editing administrative settings on the cluster controller (401) is effectively applying the settings for all nodes (101/1 - 101/n) .

Further, the content server cluster according to the present invention comprise a load balancing system. For incoming calls a load balancing unit (404) is introduced, its role being to direct incoming calls to the appropriate distribution device in the cluster in a controlled manner. The load balancing unit (404) may be a device or system designed specially for this purpose, or implemented in SIP or the like. For outgoing calls, the User Interface application residing on the cluster controller (401) acts as a load balancer, keeping track of the amount of free resources in all the CS' s in the cluster and distributing the load accordingly.

The load balancing unit (404) will process incoming video conference call requests and make a decision about which content server to handle the call. Since the desired

outputs can be generated by any of the content servers in the cluster, it is arbitrary which content server that handles the call. Consequently, the result as the user sees it will be exactly the same as if a single device was involved in the interaction.

According to one embodiment of the present invention, an H.323 gatekeeper is used as the load balancing unit (404) depicted in Figure 4. As discussed above, a gatekeeper (103) is a network device that provides addressing service for H.323 videoconference devices. Use of a gatekeeper (103) allows a videoconference device to "dial" another device using the videoconference alias rather than an IP address (which could be changed by DHCP) . In order for a gatekeeper (103) to know where to direct a call, terminals (101) and gateways (102) must register with the gatekeeper (103), informing the gatekeeper (103) of their present IP addresses and their associated aliases. The called endpoint's E.164 address may consist of an optional service prefix followed by the E.164 alias. The service prefix consists of n digits from the set of 0 to 9, * and #, where n is a predetermined number of digits. One purpose of such a service prefix might be to request access to a Gateway. The Gatekeeper may alter this address prior to sending it to the destination.

According to the present invention, the content servers (401; 106/1 - 106/n) register with its local gatekeeper (103) as gateways with one or more service prefixes and/or suffixes. By doing this, all calls directed to an address starting with one of the registered service prefixes, or ending with one of the registered service suffixes, are forwarded to said content server by the gatekeeper, regardless of the remaining digits (or alphanumeric characters) in the address. Further, all the content servers register with the gatekeeper with the same prefix

and/or suffix, e.g. 212XXXXX (that is, any address beginning with 212 and followed by 5 arbitrary digits) . Consequently, when a call request arrive at the gatekeeper with a receiver identification (called address) containing one of said prefixes or suffixes (e.g. 21255511), the gatekeeper knows that the call is to be routed to one of the registered content servers. Which content server the call request is sent to relies on available capacity and registration order. The remaining alphanumerical characters in the address can e.g. be used to tell the CS which streaming and/or archiving settings to use when streaming and/or archiving the call. The addressing problem discussed in the background section is thus removed, because all nodes have identical aliases.

Most gatekeepers implement some form of capacity indication, where devices registered as a gateway with the gatekeeper informs the gatekeeper about the amount of resources it can handle. Most gatekeepers today support an "out of resources" indication, telling the gatekeeper that a gateway has no more resources available, and therefore can not receive any further calls. If there are multiple systems registered to the gatekeeper with the same gateway prefix, the gatekeeper uses the capacity indication function to make a decision about which system will handle the incoming call. As mention above, there is a limit to how many simultaneous video conference calls the CS can handle. This number is inversely related to the complexity of the recording/streaming task. Since it is impossible to know in advance how much resources a call will demand, the CS is configured to accept only a certain number (n) of calls, wherein said number (n) typically lays in the region between 2 and 10. Therefore, according to the present invention, the Content servers are configured to send a

"out of resources" indication to the gatekeeper if its number (m) of ongoing calls is equal to its maximum allowed number (n) of calls. If one of the content servers (401; 106/1 - 106/n) , has signalled "out of resources" to the gatekeeper, the gatekeeper will route the call to another CS in the cluster with available resources. Usually, the gatekeeper will forward the call to the gateway/CS that registered first. If this gateway/CS has sent a "out of resources" message to the gatekeeper, the second registered gateway/CS is next in line, and so on. If one or more of the ongoing calls is terminated such that the number (m) of ongoing calls is less than the maximum allowed number (n) of calls in the CS, the CS sends a "free recourses" indication to the gatekeeper.

Ideally, the gatekeeper should have more detailed information about a gateway's capacity, e.g. total resources and amount of used resources at all times, not only when a gateway is out of resources.

As mentioned above, calls may also be initiated from the content server cluster, through a "create conference" option in the user interface (201) . Figure 5 is a block diagram illustrating a content server according to the present invention, set to be a cluster controller (401). When a call is initiated from the cluster, it is the load balancing unit (507) residing on a control unit (504) that controls which content server to handle the call. In order to do this, the load balancing unit (507) requires knowledge about the current capacity of all the content servers (106) in the cluster. Therefore, an application programming interface (API) call is used to gather capacity information from each device. API is a calling conventions used to allow an application program to access an operating system, as well as other system resources. There is a list

of API functions that can be called via Simple Object Access Protocol (SOAP) to interact with the content servers, e.g. make a call, end a call, get system information, etc. One of these API functions is a "GetCallCapacity" function, which for a given content server returns the maximum number of calls and the number of calls currently in progress for the given CS. When a user initiates a call from the interface (201) , the load balancing unit (507) requests all the content servers (401; 106/1 - 106/n) in the cluster for their call capacity. The content servers replies to the request with a capacity status report. Once the load balancing unit (507) has analysed the capacity information, the load balancing unit (507) can make a decision, hence spreading the load around the cluster evenly. According to the present invention, less or no load is applied to the cluster controller (401) until the nodes (106/1 - 106/n) are under full load. Since the user interface only interacts with the cluster controller (401), applying minimal load to the cluster controller (401) means that the interface can remain responsive throughout.

Figure 6 is a block diagram illustrating one exemplary embodiment of the present invention. When a conference call is recorded or streamed on one of the content servers in the cluster, the multimedia streams are converted into files (202) of preferred format and are stored on the content server handling the call. When a conference call starts, information about the recordings are written to a database (501) in the cluster controller (401), hereafter referred to as the content library. The content library contains information used by the control unit to identify where a recorded conference is stored in the cluster. Information about recordings that are being streamed live is written to the content library when the call commences, and is deleted when the call ends. Information about

recordings that can be accessed on demand are written to and kept in the content library. The information to be stored in the content library is represented by dashed lines. The control unit is configured to interact with the content library using API calls, both writing to it and reading from it. A conference initially has one recording associated with it, which contains the following information; ConferencelD, RecordingID, Title, UpdateTime and Deleted. This information is used to relate a conference to its movies and still images. A recording can have multiple movies and still images connected to it, and for each movie multiple parameters is stored in the content library, including at least RecordingID, Type, Bandwidth, DateTime, isDual etc. Further, for each still image or snapshot, the following information is stored in the content library; RecordingID, movieTime, url, Action, Description, UpdateTime and Deleted. The control unit uses this information stored in the content library for each recording, to construct Uniform Resource Locator (URL) for the user so that they can access the content and then view it using the user interface. The cluster controller can also optionally act as a recorder/streamer, but this will put extra load on the cluster controller and may have effect on its primary task, namely serving the user interface application and content library, making it slower. The cluster nodes acts only as recorders/streamers.

Figure 7 is a block diagram illustrating a second exemplary embodiment of the present invention. In this second exemplary embodiment the recording information is still stored in the content library on the cluster controller (401), but instead of storing the content on the content servers a network attached storage (NAS) is used. The NAS has its own redundancy and data protection built in. The content servers save all their recordings to the NAS, but save all the information regarding the recording to the content library as in the previous example.

The content server cluster is created by means of a setup wizard that guides the user through the process. In order to create a new cluster, the user must have administrative access to all the distribution devices designated for the new cluster. One of the distribution devices (101) is selected to operate as a cluster controller (401) . The user starts the wizard on the designated controller device (401) and supplies the installation wizard with the IP addresses of the intended nodes. For security reasons, before the controller can interact with the nodes in any way, the administrator must explicitly allow this from each node. Therefore, the setup wizard provides a link to a Graphical User Interface (GUI) for each node where the user can select whether to allow the node to be added to a cluster. This setting can only be changed by users with administrative access.

The wizard analysis the designated controller and all designated nodes to determine if the cluster can be created. Conditions that must be met are: all distribution devices must be of the same version; the nodes must have granted permission to the controller to add them to the cluster; and the licensing (number of allowed calls) must be the same. If all these conditions are met, the cluster can be successfully created.

If the cluster is created successfully, the user is presented with a setting screen identical to the settings screen for an individual device. However, now the settings apply for all systems in the cluster and will be propagated to the nodes (101/1 - 101/n) from the controller device (401) . In addition to this, all settings for call addresses and how to setup call outputs are propagated from the controller to the nodes.

The controller device (401) becomes effectively the only point of interaction to the entire cluster. Anything that happens on the controller device (401) also happens for the cluster. It is also possible to manage the cluster itself after it has been created. E.g. nodes may be added and removed at any time using the wizard or a customised version of the wizard residing on the controller device.

Users attempting to access a node via the web, using IP- address or alias, are directed to the cluster controllers interface. Administrative access is still permitted on nodes but only for the settings which apply only to the local device. These local settings may also be altered for the individual nodes from the cluster controller' s interface.