|
|
Integration of applications into the CVW is accomplished either via an application's API, the command line, or by modifications to its source code. Integration of data into the Unix CVW client is most often implemented by first defining a MIME type within the CVW and then accessing the data as it is normally accessed by other applications.
Multipoint audio and video conferencing are implemented peer to peer using vic (Video Conferencing Tool from Lawrence Berkeley Labs) and vat (Visual Audio Tool from Lawrence Berkeley Labs) via IP Multicast. The CVW server supplies multicast address information for the room-based conference to the clients transparently.
The CVW Server process is a single threaded process. Server performance is not impacted by multiple CPUs, but a faster CPU and sufficient memory to hold the entire database in RAM does greatly improve performance. The MOO server can scale up to about 300 maximum concurrent users.
A MOO has two primary components: the database (represented by and implemented in MOO-code), within which all objects and code exist to implement and populate the virtual space, and the server base, which manages incoming TCP/IP connections and executes the MOO code in the database (written in C). As new objects and features are required from the server, MOO-code is written to implement them; performance hot-spots can be recoded in C as MOO-code primitives.
The installation footprint for the CVW server is described below.
| Directory | /opt/CVWserver/ |
| Database file | CVW.db |
| Database checkpoint files | CVW.db.new, CVW.db.old.Z |
| Runtime engine (process) | moo |
| Startup script | start-mooserver |
| Log file | CVW.log |
| CVW startup script | /etc/init.d/cvw |
Requests are intended to be high-level; for example, some of the current requests used in the CVW client include room and folder contents, object overview information and multicast session establishment. By concentrating on a more semantic level, the CVW client provides interactivity and reduced bandwidth usage. Because many small sessions can be wrapped into a single TCP connection, there exists a single access point for authentication and encryption, as well as providing reduced per-connection overhead.
The Document Server is implemented using several standard components. These include: a Java servlet, a standard web server capable of running servlets, and a relational database. The servlet contains the logic for controlling access to documents. The web server provides the communication interface between the CVW client and servlet using the HTTP protocol. The relational database contains metadata about each document while the binary data for each document resides on the local file store.
The Document Server enforces single-user editing through check-out and check-in procedures. The Document Server uses MIME for file type specification, and relies upon a types database table (doctypes) to provide the mapping between the MIME type (used by the Unix CVW client) and the file extension (used by Windows). It interfaces with CVW clients and whiteboards via HTTP, thus all CVW documents are accessed as web references.
The installation footprint for the CVW Document
Server is described below.
| Directory | /opt/CVWdocserver/ |
| Administration scripts | admin/ |
| Web server | apache/ |
| Doc store | apache/htdocs |
| Java servlet | docservlet/ |
| Java servlet log file | docservlet/jserv.log |
| Java standard libraries | jserv/ |
| Relational Database | mysql/ |
| Perl tools | perl/ |
| Document server startup script | /etc/init.d/cvwds |
The CVW client's connection to the CVW server is a persistent connection. The CVW client's connection to the CVW document server is a transactional connection, and is only invoked when the user is performing document interactions. The CVW client maintains a client-side cache for document server interactions, where a local copy of documents the current user has open or checked out for editing resides. When users complete document interactions, the local document is purged from the cache.
The Unix Java CVW client uses a MIME specification in a client-side (mime-db) file to generate the list of file types that users can import into CVW. Any MIME file, such as text, image, audio, video, etc., can be imported into CVW. The Unix Java client use this file to determine which application to use when opening a CVW document for view and edit.
The Windows CVW client observes the Windows registry and file extension mapping to identify file types and helper applications used to open files and no special client-side specification files are required.
Since the CVW Document Server is based on MIME for file type specification, a customizable translation table (doctypes) resides as a database table on the document server to provide the mapping between the MIME type (used by the Unix CVW client) and the file extension (used by Windows). More information about this can be found in the Administration Guide.
The installation footprint for the Unix Java CVW Client is described below.
| Directory | /opt/CVW/ |
| Client startup script | startcvw |
| OS specific binaries (CVW, Vic, Vat, etc.) | bin/os_type/ |
| Conferencing tool binaries and libraries (Vic, Vat, mwb, admin, etc.) | modules/vic, modules/vat, modules/mwb |
| Administration tool binaries and libraries | modules/cvw-admintool |
| Client resource file | server.cvw |
| Client default preferences | master.cvwprefs |
| Client images and sounds | images/ and sounds/ |
| Client icons | images/ |
| OS specific libraries | lib/os_type/ |
| Document Mime types | lib/os_type/mime-db |
| User preferences | $HOME/.cvw/<username>/.cvwprefs |
| Client document cache | $HOME/.cvw/<username>/cache/ |
| Client document index file | $HOME/.cvw/<username>/cache/dsrvrdb |
The installation footprint for the Windows CVW Client is described below.
| Directory | C:\Program Files\MITRE\CVW\ |
| Client application | CVW.exe |
| Client resource file | Server.cvw |
| Client icons and sounds | images\ and sounds\ |
| Conferencing tool binaries and libraries | modules\vic\, modules\vat\, modules\mwb\bin and modules\mwb\library |
| Administration tool binaries and libraries | modules\cvw-admintool |
| User preferences | Users\<username>\.cvwprefs
(for Win95)
or $HOMEPATH\CVWUsers\<username>\.cvwprefs (for NT) |
| Client document cache | Users\<username>\cache\
(for Win95)
or $HOMEPATH\CVWUsers\<username>\cache (for NT) |
| Client document index file | Users\<username>\cache\dsrvdb
(for Win95)
or $HOMEPATH\CVWUsers\<username>\cache\dsrvdb (for NT) |
Vat supports a number of audio encoding schemes, varying in quality
and bandwidth requirements:
| Format | Bandwidth | CODEC |
| pcm | 78 Kbps | mu-law 8KHz PCM |
| pcm2 | 71 Kbps | mu-law 8KHz PCM |
| pcm4 | 68 Kbps | mu-law 8KHz PCM |
| dvi | 46 Kbps | Intel DVI ADPCM |
| dvi2 | 39 Kbps | Intel DVI ADPCM |
| dvi4 | 36 Kbps | Intel DVI ADPCM |
| gsm | 17 Kbps | GSM |
| lpc4 | 9 Kbps | LPC |
The CVW video conferencing capability is provided by vic, Video Conferencing Tool, developed at Lawrence Berkeley Labs. Vic source is available at LBL's web site, and binaries are available for SunOS, Solaris, Win32, IRIX, HP-UX, and LINUX. Vat supports point to point or multipoint audio conferences over an IP network. Multipoint conferences require client operating system support for IP Multicast (currently supported by Windows 95, NT 4.0, Solaris, IRIX, LINUX, etc.) and multicast routing support between the participating client's networks. The CVW server supplies multicast address information for the room-based conference to the clients, which transparently pass this information to the video tool upon invocation and when users move from room to room.
No special hardware is required to receive video, but a video capture card and camera are required for transmitting video. Vic supports bandwidth and frame rate limiting, and provides support for H.261, Sun CellB, and Xerox NV codecs. The CVW client provides a preference interface for specifying default video conferencing encoder, frame rate, and bandwidth values.
The CVW shared whiteboard capability is provided by mwb (MITRE whiteboard), developed at MITRE. Mwb provides persistent whiteboards as objects inside of the CVW. These whiteboards support drawing and image annotation by one or more parties in real time and utilizes the MCP channel between CVW clients and the CVW server to pass drawing objects between participants onto the whiteboard. Whiteboard image backdrops can be stored on the CVW Document Server, or referenced as an image location (URL) on a web server.
Mwb is implemented in Tcl/Tk and is available on SunOS 4, Solaris 2, IRIX 5&6, HP-UX 10, LINUX 2, and Win32. The same code base is used by the Tcl/Tk and Java CVW clients.
Multicasting itself is unrelated to CVW, except that it provides the underlying network infrastructure to support multipoint video and audio conferencing.
The IP multicast protocol (which is an extension to the standard IP network-level protocol) provides more efficient distribution of data over the network. IP multicast is a bandwidth reduction strategy that is in widespread but experimental stage on the Internet and on DoD networks. It provides the ability to send packets in a one-to-many fashion (e.g. streaming UDP audio/video data) through the use of IP class D address extensions (224.0.0.0 to 239.255.255.255). The address range of 224.2.*.* is the range used by the MBone (Internet Multicast Backbone) users for audio and video conferencing sessions. Sessions are assigned multicast addresses, and users join the session by subscribing to the multicast address with their client software. Session participants may join and leave the session at any time if they know the multicast address for the session.
Network routers that can route multicast traffic are required. If multicast-capable routers are unavailable, a multicast tunneling scheme can be employed to forward multicast packets among participating subnets by encapsulating the multicast packets inside regular IP packets. Multicast tunnels are made up of one tunnel endpoint (workstation running the mrouted software) on each subnet that needs to be interconnected.
The audio and video conferencing tools integrated with CVW, LBL vic and vat, use IP multicast as supporting network infrastructure for mulitpoint audio and video conferencing. The CVW Server does not participate in the multicast network. The role of the CVW Server is to provide a unique multicast address for each CVW room to the CVW client, so that the client can pass the address to the conferencing tools.
A complete multicast address is made up of 4 octets. In the CVW, the first two octets (i.e., 226.10) make up the base address for the CVW Server. This base address is provided by the CVW administrator in the CVW System Settings. The second 2 octets make up the room address in which the audio and video conference is being held. These are automatically sent by the CVW Server to the CVW client when a user walks into a room. Each room in the CVW has a unique predefined room address which is used by the CVW Server in managing the multicast conferences for that room.
What makes the full multicast address unique for each room across all CVW Servers is the base address provided by the CVW Server administrator. It is important that the base address for each new CVW Server on the same network be unique so that the multicast traffic is kept separated from server to server. For example, if two CVW Servers on the same network shared the same base address, users in the rooms corresponding to the location of room 101 on each server would appear in each other's audio and video conferences because they would be sharing the same multicast address.
Last update: 13 October 2000
This work, including the source code, documentation and
related data, is placed into the public domain.
The original author is
The MITRE Corporation.
THIS SOFTWARE IS PROVIDED "AS IS" AND WITHOUT ANY
WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, AS TO THIS
SOFTWARE, OR TO THE ACCURACY, CAPABILITY, EFFICIENCY,
OR FUNCTIONING OF THIS SOFTWARE OR ANY DERIVATIVE WORK
OR OTHER ITEM MADE, USED, OR SOLD THAT INCORPORATES
THIS SOFTWARE. THE AUTHOR OF THIS SOFTWARE ASSUMES NO
RESPONSIBILITY FOR ANY CONSEQUENCE RESULTING FROM THE
USE, MODIFICATION, OR REDISTRIBUTION OF THIS SOFTWARE.
Information in this document is subject to change without notice.
Other products and companies referred to herein are trademarks or registered
trademarks of their respective companies or mark holders.