Home Products Solution Examples Support Company

Solution Examples

Since 2003, GraphStream has delivered scalable data systems to many hundreds of client sites worldwide. Our high rate of repeat business is an indicator of strong customer satisfaction.

The following sections provide some examples of how GraphStream works together with partners and clients in multiple fields to improve the efficiency of commercial and technical IT applications:

Solution Examples: Healthcare

GraphStream offers a range of solutions for radically improving the efficiency of clinical-information management and access within single-site and multi-site healthcare enterprises. Here are two examples:

Calgary Scientific offers software packages that provide entirely new capabilities and enhanced workflows for interacting with clinical information. The FDA-cleared ResolutionMD suite provides advanced clinical-information acess, visualization, and analysis capabilities, with secure, regulation-compliant access from any location.
ResolutionMD uses servers to render visualizations of clinical data retrieved from one or more archives using industry-standard protocols. These visualizations are then sent to one or more client devices that enable user interaction. A key element of the data-security compliance model is that only the visualizations, and not the clinical data, are sent to client devices.

ResolutionMD supports access from a range of client devices, including high-end diagnostic workstations; wireless mobile platforms such as Apple IOS (e.g. iPad, iPhone) and Android; and any other system equipped with a Web browser.

ResolutionMD delivers superior low-latency interactivity on client devices, and allows multiple users to work collaboratively in a shared visual workspace. Each participant in a collaboration session may join from any location worldwide using any supported client device, and is provided with a uniquely-colored cursor that is visible to all participants.

GraphStream offers services and infrastructure for ResolutionMD deployments. We offer multiple options for the required infrastructure, including choices of on-site or off-site, and dedicated or cloud-based.
The clinical information that a healthcare enterprise must manage may include imaging studies from radiology, cardiology, oncology, and other specialties; digital-pathology images; video recordings from endoscopy; EEG and EKG data from sleep studies; laboratory-test reports; document scans; and many other data types. In many organizations, this rapidly growing mass of information is scattered across numerous specialized department-level hardware systems that use myriad vendor-proprietary data formats and inefficient, outdated storage technologies.
GraphStream offers solutions that enable organizations to take control of their clinical information; migrate it into modern enterprise-IT infrastructure; store it using open industry-standard data formats; apply uniform best-practices for regulatory compliance, backup, data integrity, and security; and provide efficient, secure access to users via industry-standard interfaces.

For enterprise clinical-data archive implementations, GraphStream offers both services and infrastructure. Available services include system design and planning; installation and configuration; and data migration/ingest. As with ResolutionMD mentioned above, we offer multiple options for the required infrastructure, including choices of on-site or off-site, and dedicated or cloud-based.

Solution Examples: Energy

GraphStream works with clients in the global upstream Oil&Gas industry to deploy a variety of custom-configured systems that dramatically improve the efficiency of specific applications. Here are some examples:

Energy exploration and production companies have volumes of highly proprietary mission-critical shared data that need to be accessed securely and simultaneously for both reading and writing from numerous staffed and unstaffed sites all over the world. Some of these sites are extremely remote (e.g. offshore platforms) and resource-constrained, with little or no local IT support capability.
There are many ways to provide the infrastructure needed for this application. Most involve either high-cost, low-volume, specialized vendor-proprietary software and hardware components, or complex and expensive combinations of general-purpose enterprise-IT products from multiple suppliers, including Network-Attached Storage units, local disk-to-disk or tape backup systems, hardware appliances for security and for WAN optimization, and software licenses for a variety of extra-cost capabilities.

GraphStream offers an alternative solution for this application, based on the Nasuni Data Continuity Services, that provides superior capability while radically reducing total cost, complexity, space/power/cooling footprint, and maintenance requirements. This solution scales easily to any number of sites and any quantity of shared data.

Each site deploys a single virtual or physical appliance, which can take the form of a small rackmount server, a compact embedded system, or even a notebook computer. This system provides access via CIFS and NFS protocols to one or more high-availability cloud-based shared storage volumes, including standard enterprise-IT access controls, instantaneous elastic capacity scaling, encryption, file-level versioning, deduplication, thin provisioning, off-site data protection, disaster recovery with unprecedented performance, caching for local access performance, and WAN optimization. All of this capability covering all sites is provided for a single annual fee based only on the amount of data actually being stored. The total cost of this solution is a small fraction of what alternatives cost.
Companies that offer seismic-data processing services seek advantage in a highly competitive market by integrating and optimizing complete scalable systems that combine proprietary processing algorithms and software with efficient networking, processing, and storage infrastructure based on Commercial Off-The-Shelf (COTS) components.
GraphStream is currently providing our seismic-processing clients with highly scalable systems that incorporate a wide range of software and hardware components, including Bright Computing cluster-management software, Mellanox InfiniBand-FDR 56Gbps network fabrics, Intel Xeon CPUs, NVIDIA Tesla GPUs for processing acceleration, and solid-state storage systems for I/O acceleration.

GraphStream has been deploying commercial scalable technical-computing systems with InfiniBand networking and per-node GPU acceleration since 2003, far longer than any other supplier. GraphStream's seismic-processing clients benefit from our superior depth of experience with this type of infrastructure, as we work with them to plan and execute intelligent strategies for phased incremental deployments of additional infrastructure capacity to meet their growing needs.

Cost per unit of networking, processing, and storage capacity is continuing to drop rapidly, with frequent introductions of new, dramatically more efficient component designs. For future scalable-system designs, higher-speed and lower-latency Ethernet networking will compete against InfiniBand, and upcoming single-chip processors will compete against multi-chip GPU+GPU systems. Additionally, seismic-processing providers have traditionally acquired capacity by purchasing and leasing hardware, but ongoing improvements in the cost-efficiency and security of cloud-based infrastructure are leading to increasing interest in capacity-on-demand models. GraphStream will continue to work together with our seismic-processing clients to navigate these technology transitions successfully.

Solution Example: Manufacturing

GraphStream works with partners to deploy scalable interactive systems for computer-aided design and manufacturing. These systems use virtualization technologies to share datacenter-based pools of networking, processing, and storage capacity among multiple concurrent interactive users. This sharing enables superior flexibility and efficiency in resource utilization, and can span multiple interaction modalities including:

Team rooms with immersive multi-channel video and audio output systems, and a variety of input devices.
Desktop workstations with multiple high-resolution displays.
Mobile wireless input/display units, including tablets and smartphones.

Solution Example: Financial Services

GraphStream's financial-services clients use advanced scalable data networking, processing, and storage systems in real-time trading applications. Key competitive attributes of these systems include:

Superior space and power efficiency, which helps to minimize latency by enabling installation of systems in close physical proximity to exchange-operated data plants.

Lower network latency, using leading-edge Ethernet and InfiniBand networking components from Arista and Mellanox.

Higher processing throughput, via a combination of Intel Xeon CPUs and NVIDIA Tesla GPUs.
Superior storage latency and throughput, enhanced in some cases by Fusion-io solid-state storage modules.

Solution Example: Government/Defense

The U.S. Department of Energy (DoE) uses some of the world's highest-capacity supercomputing systems to run physical simulations. These simulations produce terabyte-scale datasets as output. DoE continues to seek more-efficient ways to produce visualizations of these datasets.

GraphStream responded to this DoE requirement by working together with Intel, NVIDIA, and Mellanox to deliver the world's first commercially integrated scalable technical-computing systems with GPU acceleration in each processing node and InfiniBand networking.

GraphStream has deployed a number of the world's largest systems of this type to DoE National Laboratories: Argonne, Idaho, Lawrence Livermore, Los Alamos, Oak Ridge, and Sandia. Like many of GraphStream's clients, DoE has entrusted GraphStream with mission-critical, multi-million-dollar integration projects, and GraphStream has delivered systems with leading-edge performance and efficiency, on-schedule and on-budget.

GraphStream 18-cabinet, 257-server scalable visualization system at Lawrence Livermore National Laboratory