Shantanu Mitra, PhD.
Johan Pirot
Claude Sandroff
TeraBurst Networks
Sharing visualization resources across the enterprise
Advances in hardware and software capabilities for scientific visualization have resulted in a growing interest in its utilization as an effective competitive tool for enterprises in various market segments. High-quality immersive visualization can enhance the understanding, interpretation, and modeling of large amounts of information in the oil and gas industry, and several visualization theaters have recently been installed.
In the area of oil exploration, one of the primary objectives of the activity in the field is to maximize the effectiveness of the drilling location for more efficient production cycles. This requires understanding and interpreting two closely connected systems:
- The subsurface reservoir
- The surface recovery, property separation, storage, and processing.
Significant advances have been made over the past decade in the production cycle, but none have been more so than subsurface characterization modeling. The 3D transient simulation tools used for subsurface modeling involve terabytes of time-elapsed multi-dimensional seismic field data. Tools for storing and displaying these data through immersive visualization exist at some of the major international installations and are considered an indispensable tool in today's environment. Enterprises are looking to maximize their investment in these multi-million-dollar visualization installations and are actively considering ways to make the resource available to a larger group of users.
It is often critical that sub-surface characterization models used to investigate potential drilling sites be viewed, shared, and worked collaboratively by as many scientific experts as needed. This can be a slow and expensive process, as the experts need to travel to the visualization theaters where the data reside – often in geographically disparate locations – resulting in travel fatigue and inefficiency. In some cases, several dozen experts need to be transported to the data. In some cases, nations with drilling and exploration operations place embargos on the export of data from geological exploration, making it imperative for the team to travel to the data.
A collaborative visualization solution could address this need. The requirements include:
- Real-time connectivity over a long-range network with local or global reach
- 3D stereoscopic image capability
- Audio links or video conferencing capabilities.
In some cases, visualized images need to be presented and distributed to several remote locations within the enterprise. Applications include training, presentations, project updates, and simply sharing visualized images. Require-ments for this group of remote visual- ization applications also include real-time connectivity over a network with local or global reach and 3D stereoscopic image capability for viewing at the remote location.
Distributed visualization
Today, enterprises can consider using public networks to connect visualization installations. Available solutions, however, have several issues associated with them. They include, depending on the architecture of the specific solution:
- Expensive databases (hardware and software) at multiple locations
- Large datasets that cannot be efficiently transferred
- A larger number of expensive software licenses at multiple locations
- The nature of the bandwidth between locations for real-time, low-latency connectivity.
It is undesirable then to transport the actual data. Sol-utions based on the transport of the images generated from the data are more attractive.
The current status of distributed visualization lacks connection solutions with low delay and global reach.
null
Typical packet-based services used with the current solutions, e.g., IP network services or T-1 lines from the local exchange carrier, are capable of providing throughputs in the hundreds of kilobits/second to 1.5 Mbps. High-resolution graphics in the visualization centers typically utilize bandwidths in the gigabit/second range.
For example, consider a stereoscopic 3D imaging application with display resolutions of 1024 x 768 or 1280 x 1024 pixels. For the stereoscopic capability, every image frame is sent twice – once for the left eye and once for the right, at refresh rates between 30 to 56 Hz for each eye, resulting in total frame-refresh rates up to 112 Hz.
For SXGA resolutions at 96 Hz, this results in an uncompressed data rate of ~ 3 Gbps (1280 x 1024 x 96 x 24). An IP network cannot provide this type of bandwidth, and compression techniques must be used.
Additionally, packet networks typically operate on a "best effort" basis, without any guarantee of data arriving in the proper sequence, resulting in buffering requirements and additional delay. Further, for a 3D image undergoing compression, random packet loss through the network can result in quality, e.g., the exact same pixel must be "dropped" in both the left and right frames to prevent artifacts.
In the case of multi-screen displays, unpredictable packet delivery can also cause loss of synchronization between the screens.
As can be seen in the figure, no satisfactory solution exists for the problem of real-time distributed visualization between geographically disparate visualization theaters – especially if more than a few kilometers separate these theaters. Current attempts suffer from distance limitations, are not truly real-time applications, or are not capable of high-resolution 3D stereoscopic graphics.
The basic terminal functionality required for the graphics/optics converter for telecommunications networks exceeds the bandwidth on common T-1 lines.
null
Optical networking
A new wide-area visualization solution (WAVS) has recently been developed. This approach is also based on the concept of transporting the visualized image and not duplicating databases or software licenses, in real-time, over global networks with a 3D stereoscopic capability. The WAVS solution, however, makes use of synchronous optical networking (Sonet), a well established standard used in telecommunications networks across the US. Sonet, along with its closely related international version, synchronous digital hierarchy (SDH), forms the basis for the global telecommunications network.
The optical converter inputs include graphics (RGB), H-sync, V-sync, and stereo emitter signals from the computer, along with audio inputs, keyboard/mouse inputs, and a 10/100 LAN port input. On the output side is a single optical Sonet/SDH signal. Different versions of this system have been developed and, based on a 2.5 Gbps or 155 Mbps Sonet/SDH signal, results in uncompressed or compressed graphics.
To be an effective tool for remote collaboration, the system also has an optical return path from the remote site, carrying the same signals (graphics, control, data, audio), with the ability to have the "remote" keyboard and mouse control the "local" data.
As Sonet/SDH is a widespread transmission standard, it is possible to transport the digitized information over a public network irrespective of the different service providers involved in offering the end-to-end connectivity.
Additionally, Sonet/SDH is a transmission standard that is circuit-based and not packet-based. This implies that the image data are carried in the correct sequence and reproduced at the remote site in that sequence. Data packets on a packet-switched network are not guaranteed to arrive in the correct sequence, thereby introducing latency as the image needs to be buffered and rebuilt in the proper sequence. Additionally, data packets can also be lost without recourse to efficient recovery, potentially resulting in serious artifacts in the images at the remote site.
Some beneficial features of a distributed visualization solution based on this optical networking technology are:
- 3D stereoscopic graphics capability: For large composite images that are displayed on multiple blended screens, the WAVS solution involves multiple systems that are frame-locked to ensure that they must also be capable of maintaining the frame sequence across all the screens. The use of a circuit-based optical networking protocol, such as Sonet/SDH, ensures this
- Real-time connectivity with global reach: To make interaction meaningful, there has to be minimal delay in the viewed image. This is not always possible with a packet-based network due to packet loss, rerouting of individual packets, and buffering at the destination. The use of synchronous optical networks avoids this issue and guarantees arrival of data in the right sequence at the destination. Furthermore, the use of Sonet/SDH protocols guarantees global connectivity. The WAVS solution has been tested over 18,000 km with latencies of 100 milliseconds
- Fully interactive experience: This approach also allows synchronization between audio and video, along with control signals and a 10 Mbps data channel. All the signals are carried over a single wavelength of light in the Sonet/SDH network. This allows remote control of the data and supports third-party video conferencing applications
- Security: Since the WAVS solution only transports images in real-time, there are no security concerns associated with transporting data. Furthermore, the public network never actually gets attached to the computer hosting the data. Images are only transported when a session is in progress, with a similar system being required at the remote end. Additionally, circuit-based technologies like Sonet/SDH tend to be more secure than packet-based networking technology
- Broadcast capabilities: Because the WAVS solution includes all video, audio, and control information in a single wavelength, this wavelength can easily be split, replicated, and delivered to multiple destinations. A broadcast from one source to multiple destinations can be established, and multiple parties in different locations can all work together on the same session at the same time.
Graphics images from a single dataset (computer), along with audio, keyboard/mouse control, and LAN data can be displayed 40 km away without additional networking equipment.
Global connectivity over public networks can link widely spaced locations.
null
New applications
The graphics/optics converters can be connected to each other directly or through any telecom-ready optical transmission equipment like Dense Wavelength Division Multiplexing (DWDM) gear for increasing the distance of transmission. With a direct link, the systems are capable of a separation of up to 40 km on standard single-mode fiber. For longer distances over private fiber, standard DWDM equipment operating at either OC-48 or OC-3 (depending on the graphics converter system) can be used to reach distances of up to 400 km. Only one pair of fibers is required for a bi-directional link.
While the use of Sonet/SDH optical networking brings certain benefits to the configuration above, the true advantage can be seen with the next example. Here, the public telecommunications network, via an OC-3 Sonet/SDH service provides connectivity between Bakersfield, California, Houston, Texas, and Scotland. Typically, bandwidth is provisioned through the optical network to the network access point closest to the end-user facilities. Connectivity between international sites is also possible through OC-3 wavelength services offered by the international carriers. Final, "last mile" connectivity is obtained either from private fiber or local exchange carriers. The optical switch in the figure can be used to select between two different sources and is compatible with standard Sonet/SDH signals.
Distributing visualization resources across global enterprises is now possible with a new synchronous optical networking based hardware solution. The use of this networking technology allows a wide area visualization solution with distinct advantages over existing solutions. A new set of innovative applications can be enabled by this technology.
For information, contact Johan Pirot, president, TeraBurst Networks: (408) 400-4140; email: [email protected].
