ConnectX VPI Single/Dual-Port Adapters with Virtual Protocol Interconnect

Transcription

1 ADAPTER CARDS ConnectX VPI Single/Dual-Port Adapters with Virtual Protocol Interconnect ConnectX adapter cards with Virtual Protocol Interconnect (VPI) provide the highest performing and most flexible interconnect solution for Enterprise Data Centers, High-Performance Computing, and Embedded environments. Clustered data bases, parallelized applications, transactional services and high-performance embedded I/O applications will achieve significant performance improvements resulting in reduced completion time and lower cost per operation. ConnectX with VPI also simplifies network deployment by consolidating cables and enhancing performance in virtualized server environments. Virtual Protocol Interconnect VPI-enabled adapters make it possible for any standard networking, clustering, storage, and management protocol to seamlessly operate over any converged network leveraging a consolidated software stack. With auto-sense capability, each ConnectX port can identify and operate on InfiniBand, Ethernet, or Data Center Ethernet (DCE) fabrics. ConnectX with VPI simplifies I/O system design and makes it easier for IT managers to deploy infrastructure that meets the challenges of a dynamic data center. World-Class Performance Over InfiniBand ConnectX delivers low-latency and high-bandwidth for performance-driven server and storage clustering applications. These applications will benefit from the reliable transport connections and advanced multicast support offered by ConnectX. Network protocol processing and data movement overhead such as InfiniBand RDMA and Send/Receive semantics are completed in the adapter without CPU intervention. Servers supporting PCI Express 2.0 with 5GT/s will be able to take advantage of 40Gb/s InfiniBand, balancing the I/O requirement of these high-end servers. World-Class Performance Over Data Center Ethernet ConnectX delivers similar low-latency and high-bandwidth performance leveraging Ethernet with DCE support. Low Latency Ethernet (LLE) provides efficient RDMA transport over Layer 2 Ethernet with IEEE 802.1q VLAN tagging and Ethernet Priority-based Flow Control (PFC). The LLE software stack maintains existing and future application compatibility for bandwidth and latency sensitive clustering applications. With link-level interoperability in existing Ethernet infrastructure, Network Administrators can leverage existing data center fabric management solutions. TCP/UDP/IP Acceleration Applications utilizing TCP/UDP/IP transport can achieve industry-leading throughput over InfiniBand or 10GigE. The hardware-based stateless offload engines in ConnectX reduce the CPU overhead of IP packet transport, freeing more processor cycles to work on the application. I/O Virtualization ConnectX support for hardware-based I/O virtualization provides dedicated adapter resources and guaranteed isolation and protection for virtual machines (VM) within the server. I/O virtu- ConnectX Adapter Cards Single- and Dual-Port QSFP Connectors Shown BENEFITS One adapter for InfiniBand, 10Gig Ethernet or Data Center Ethernet fabrics World-class cluster performance High-performance networking and storage access Guaranteed bandwidth and low-latency services Reliable transport I/O consolidation Virtualization acceleration Scales to tens-of-thousands of nodes KEY FEATURES Virtual Protocol Interconnect 1us MPI ping latency Selectable 10, 20, or 40Gb/s InfiniBand or 10GigE per port Single- and Dual-Port options available PCI Express 2.0 (up to 5GT/s) CPU offload of transport operations End-to-end QoS and congestion control Hardware-based I/O virtualization TCP/UDP/IP stateless offload Fibre Channel encapsulation (FCoIB or FCoE)

2 ConnectX VPI Single/Dual-Port Adapters with Virtual Protocol Interconnect 2 alization with ConnectX gives data center managers better server utilization and LAN and SAN unification while reducing cost, power, and cable complexity. Storage Accelerated A consolidated compute and storage network achieves significant costperformance advantages over multi-fabric networks. Standard block and file access protocols leveraging InfiniBand RDMA result in high-performance storage access. Fibre Channel frame encapsulation (FCoIB or FCoE) and hardware offloads enable simple connectivity to Fibre Channel SANs. Software Support All Mellanox adapter cards are compatible with TCP/IP and OpenFabrics-based RDMA protocols and software. They are also compatible with InfiniBand and cluster management software available from OEMs. The adapter cards are compatible with major operating system distributions. Adapter Cards FEATURE SUMMARY INFINIBAND IBTA Specification compliant 10, 20, or 40Gb/s per port RDMA, Send/Receive semantics Hardware-based congestion control Atomic operations 16 million I/O channels 256 to 4Kbyte MTU, 1Gbyte messages 9 virtual lanes: 8 data + 1 management ENHANCED INFINIBAND Hardware-based reliable transport Hardware-based reliable multicast Extended Reliable Connected transport Enhanced Atomic operations Fine grained end-to-end QoS ETHERNET IEEE 802.3ae 10 Gigabit Ethernet IEEE 802.3ak 10GBASE-CX4 IEEE 802.3ad Link Aggregation and Failover IEEE 802.3x Pause IEEE 802.1Q,.1p VLAN tags and priority Multicast Jumbo frame support (10KB) 128 MAC/VLAN addresses per port Class Based Flow Control / Per Priority Pause HARDWARE-BASED I/O VIRTUALIZATION Single Root IOV Address translation and protection Multiple queues per virtual machine VMware NetQueue support PCISIG IOV compliant ADDITIONAL CPU OFFLOADS TCP/UDP/IP stateless offload Intelligent interrupt coalescence Compliant to Microsoft RSS and NetDMA STORAGE SUPPORT T10-compliant Data Integrity Field support Fibre Channel over InfiniBand or Ethernet COMPATIBIITY CONNECTIVITY Interoperable with InfiniBand or 10GigE switches microgigacn or QSFP connectors 20m+ (10Gb/s), 10m+ (20Gb/s) or 7m+ (40Gb/s) of pasive copper cable External optical media adapter and active cable support MANAGEMENT AND TOOLS InfiniBand OpenSM Interoperable with third-party subnet managers Firmware and debug tools (MFT, IBDIAG) Ethernet MIB, MIB-II, MIB-II Extensions, RMON, RMON 2 Configuration and diagnostic tools OPERATING SYSTEMS/DISTRIBUTIONS Novell SLES, Red Hat Enterprise Linux (RHEL), Fedora, and other Linux distributions Microsoft Windows Server 2003/2008/CCS 2003 OpenFabrics Enterprise Distribution (OFED) OpenFabrics Windows Distribution (WinOF) VMware ESX Server 3.5, Citrix XenServer 4.1 PROTOCOL SUPPORT Open MPI, OSU MVAPICH, HP MPI, Intel MPI, MS MPI, Scali MPI TCP/UDP, IPoIB, SDP, RDS SRP, iser, NFS RDMA, FCoIB, FCoE udapl Ordering Part Number Port Host Bus Power Typ Dimensions w/o Brackets MHGH28-XTC Dual 4X 20Gb/s InfiniBand or 10GigE PCIe GT/s 11.0W 13.6cm x 6.4cm MHGH29-XTC Dual 4X 20Gb/s InfiniBand or 10GigE PCIe GT/s 11.6W 13.6cm x 6.4cm MHJH29-XTC Dual 4X 40Gb/s InfiniBand or 10GigE PCIe GT/s 12.2W 13.6cm x 6.4cm MHRH29-XTC Dual 4X QSFP 20Gb/s InfiniBand PCIe GT/s 11.6W 16.8cm x 6.4cm MHQH29-XTC Dual 4X QSFP 40Gb/s InfiniBand PCIe GT/s 12.2W 16.8cm x 6.4cm Single-port options also available Oakmead Parkway, Suite 100, Sunnyvale, CA Tel: Fax: Copyright Mellanox Technologies. All rights reserved. Mellanox, ConnectX, InfiniBlast, InfiniBridge, InfiniHost, InfiniRISC, InfiniScale, and InfiniPCI are registered trademarks of Mellanox Technologies, Ltd. Virtual Protocol Interconnect and BridgeX are trademarks of Mellanox Technologies, Ltd. All other trademarks are property of their respective owners. 2770PB Rev 2.5

3 CABLES Copper Cables Supporting data rates up to 40Gb/s Mellanox cables are a cost-effective solution for connecting InfiniBand adapters, extending the benefits of Mellanox s high-performance InfiniBand adapters throughout the network. These cables meet or exceed IBTA and IEEE standards, and have been rigorously tested with all Mellanox products. Mellanox passive copper cables provide robust connections up to seven meters in length and require no power from the connector. Mellanox active cables are a thinner alternative to the passive six and eight meter cables, and are available up to 12 meters in length. All cables have easy-to-read serial numbers printed at each end to facilitate installation and network debug. Enterprise Data Centers, High-Performance Computing, and Embedded installations are assured the highest interconnect performance when using Mellanox adapters and cables. Cables Ordering Part Number Length (m) Wire AWG Cable Type MCC4Q30C-00A QSFP MCC4Q30C QSFP MCC4Q26C QSFP MCC4Q26C QSFP MCC4Q26C QSFP MCC4Q26C QSFP MCC4Q26C QSFP MCC4Q26C QSFP MCC4N26C Hybrid QSFP-CX4 MCC4N26C Hybrid QSFP-CX4 MCC4N26C Hybrid QSFP-CX4 MCC4R26C Active QSFP MCC4R28C Active QSFP MCC4R26C Active QSFP MCC4R26C Active QSFP BENEFITS Industry-leading price/performance Certified performance on Mellanox-based products KEY FEATURES Compliant to the InfiniBand Architecture specification Supports up to 40Gb/s bandwidth Superior signal integrity Bit Error Rate (BER) less than Low crosstalk Low insertion loss Matched impedance Secure latching mechanism Serial numbers printed at each end RoHS-5 compliant 1-year warranty SPECIFICATIONS Passive cables - Maximum reach: 7 meters (24 AWG) Active cables - Maximum reach: 12 meters (26 AWG) - Maximum current: 400 mamps - Power supply: 3.3 Volts 350 Oakmead Parkway, Suite 100, Sunnyvale, CA Tel: Fax: Copyright Mellanox Technologies. All rights reserved. Mellanox, ConnectX, InfiniBlast, InfiniBridge, InfiniHost, InfiniRISC, InfiniScale, and InfiniPCI are registered trademarks of Mellanox Technologies, Ltd. Virtual Protocol Interconnect and BridgeX are trademarks of Mellanox Technologies, Ltd. All other trademarks are property of their respective owners. 2961PB Rev 1.1

4 SWITCH SYSTEM MTS port 20 and 40Gb/s InfiniBand Switch System MTS3600 switch systems provide the highest-performing fabric solution by delivering high-bandwidth and low-latency to Enterprise Data Centers, High-Performance Computing and Embedded environments. Networks built with MTS3600 systems can carry converged traffic with the combination of assured bandwidth and granular quality of service. World Class Density and Scalability Built with Mellanox s 4th generation InfiniScale IV InfiniBand switch device, the MTS3600 provides up to 40Gb/s full bisectional bandwidth per port. With 36 ports, these systems are among the densest switching systems available. These stand-alone switches are an ideal choice for topof-rack leaf connectivity or for building small to medium sized clusters. Sustained Network Performance The MTS3600 supports adaptive routing as well as static routing to reduce or eliminate congestion situations. Adaptive routing dynamically and automatically re-routes traffic to alleviate congested ports. Static routing, meanwhile, has been shown to produce superior results in networks where traffic patterns are more predictable. In situations where switching contention is unavoidable, such as many-to-one traffic patterns, the MTS3600 supports IBTA 1.2 congestion control mechanisms. The switch systems work in conjunction with ConnectX InfiniBand HCAs to restrict the traffic causing congestion, ensuring high-bandwidth and low-latency to all other flows. Whether used for parallel computation or as a converged fabric, the combination of high-bandwidth, adaptive routing, and congestion control provide the industry s best traffic-carrying capacity. Utility Computing Virtual partitioning of a cluster enables efficient use of all of its computing resources. Allocating only the compute power that each client needs enables more clients on the cluster at one time. Clusters built on MTS3600 systems can run up to six multiple subnets, securely segregating client processes while ensuring the highest productivity of the cluster. High Availability The MTS3600 systems deliver director-class availability required for mission-critical applications environments. These systems support an optional redundant hot-swappable power supply and a hot-swappable fan unit to help eliminate down time. Easy Management The MTS3600 systems are easily managed through any IBTA compliant subnet manager. Port configuration and data paths can be set up automatically or customized to meet the needs of the application. Firmware can be updated in-band for simple network maintenance. 1U 36-port InfiniBand Switch with QSFP connectors BENEFITS Industry-leading stand alone switch platform in performance, power, and density Wirespeed InfiniBand switch platform up to 40Gb/s per port High-bandwidth, low-latency fabric for compute-intensive applications KEY FEATURES 2.88Tb/s switching capacity Adaptive routing Congestion control Quality of Service enforcement Up to 6 virtual subnets Temperature sensors and voltage monitors INFINIBAND IBTA Specification 1.2 compliant Integrated subnet manager agent Hardware-based congestion control Adaptive routing Fine grained end-to-end QoS 2.88Tb/s total switching bandwidth 256 to 4Kbyte MTU 9 virtual lanes: 8 data + 1 management 48K entry linear forwarding data base MANAGEMENT Supports Open SM or third-party subnet managers Up to 6 virtual subnets with dynamic port allocation Port mirroring Diagnostic and debug tools

5 MTS port 20 and 40Gb/s InfiniBand Switch System 2 Mellanox Advantage Mellanox is the leading supplier of industry standard InfiniBand adapters and switch products. Our products have been deployed in clusters scaling to thousands of nodes and are being deployed end-to-end in data centers and Top500 systems around the world. SPECIFICATIONS 36 InfiniBand ports - Up to 40Gb/s - Up to 3.5W per port for active cable or fiber module support Full bisectional bandwidth to all ports IBTA 1.2 compliant Dual redundant auto-sensing 110/220VAC power supplies Hot-swappable fan trays Port and system status LED indicators RoHS-5 compliant Rack-mountable, 1U 1-year warranty HARDWARE INFINIBAND SWITCH 36 ports Up to 40Gb/s per port Fully non-blocking architecture CONNECTORS AND CABLING QSFP connectors Passive copper cable Active copper and fiber cables Fiber media adapters INDICATORS Per port status LEDs: Link, Activity System status LEDs: System, fans, power supplies DIMENSIONS AND WEIGHT 1.73H x 17.5W x 23.8D inches 22 lbs (1 power supply) 24.5 lbs (2 power supplies) POWER SUPPLY Dual redundant slots Hot plug operation Input range: VAC Frequency: 47-63Hz, single phase AC MAXIMUM POWER CONSUMPTION 316W (MTS3600R-1UNC) - Dissipated power: 190W - Power through QSFP: 126W 326W (MTS3600Q-1UNC) - Dissipated power: 200W - Power through QSFP: 126W COMPLIANCE SAFETY US/Canada: ctuvus EU: IEC60950 International: CB EMC (EMISSIONS) USA: FCC, Class A Canada: ICES, Class A EU: EN55022, Class A EU: EN55024, Class A EU: EN , Class A EU: EN , Class A Japan: VCCI, Class A ENVIRONMENTAL EU: IEC : Random Vibration EU: IEC : Shocks, Type I / II EU: IEC : Fall Test ACOUSTIC ISO 7779 ETS OPERATING CONDITIONS Operating temperature: 10 to 45 C Humidity: 10-90% non-condensing Ordering Part Number MTS3600Q-1UNC RoHS R5 Reserved Optional management U= unmanaged A = Celeron-based management Number of power supplies InfiniBand port speed and port connector R = 10, 20Gb/s QSFP Q = 10, 20, 40Gb/s QSFP InfiniScale IV 36-port Switch System Switch System Mellanox Technologies 350 Oakmead Parkway, Suite 100, Sunnyvale, CA Tel: Fax: Copyright Mellanox Technologies. All rights reserved. Mellanox, ConnectX, InfiniBlast, InfiniBridge, InfiniHost, InfiniRISC, InfiniScale, and InfiniPCI are registered trademarks of Mellanox Technologies, Ltd. Virtual Protocol Interconnect and BridgeX are trademarks of Mellanox Technologies, Ltd. All other trademarks are property of their respective owners. 2938PB Rev 1.0

6 Platform Cluster Manager Benefits: Highlights: Easily deploy open clusters Simplify cluster management Update compute nodes without re-installing Perform risk-free, trial software installations Change node personalities with ease Deploy your choice of Linux OS environments The most advanced cluster management tools Integrated workload management Intuitive web-interface Optimized for performance Certified & supported Clusters Made Simple A better way to manage open clusters Key capability Clusters installed quickly and easily Easily install add-on software Available NVIDIA CUDATM SDK Easy to use web interface Update software with no re-installing Perform repository snapshots Open Source Solution Platform Cluster Manager Platform Cluster Manager (PCM) represents a re-think of how open source clusters are deployed and managed. Based on the open source Project Kusu, Platform Cluster Manager is a pre-integrated, vendor certified solution enabling the consistent delivery of scaled-out application clusters. Built using open source components, it includes all the tools required to deploy, run and manage clusters with unprecedented ease. Using Platform Cluster Manager, cluster administrators can easily perform operations that are simply not possible with less advanced cluster management solutions. Change node personalities any time Integrated workload manager User-friendly job submission portal Built-in monitoring, alerting & reporting Multiple OS environments Vendor optimized 24/7 Enterprise support Easy to use web interface The web interface makes operations like installing and maintaining clusters a snap. Cluster administrators will appreciate the built-in alerting, reporting and management features and the ease with which software configurations can be managed and deployed without ever leaving the Platform Management Console (PMC) graphical interface. Cluster users can take advantage of the web portal and the open workload manager, Platform Lava (Platform LSF compatible), that allows them to submit, monitor and manage their own cluster workloads without ever touching the command line. Ideal for academic and commercial applications alike, the intuitive web interface makes job submission interfaces self-documenting reducing training requirements and error rates and reducing the support burden on busy cluster administrators.

7 Platform Cluster Manager provides an intuitive extensible web interface allowing administrators to manage the cluster without learning complex command line syntax. Cluster users will appreciate the easy to use web interface for submitting and monitoring simulations and cluster applications. Manage cluster software as installable components Update compute nodes on the fly As clusters inevitably get busy and user communities grow, scheduling downtime even in small environments can be a challenge. This is especially true in environments with multiple users, or those supporting long-running or parallel workloads. While other cluster management solutions require that nodes be re-installed every time a minor software update is made, Platform Cluster Manager provides a powerful cluster file manager (CFM). For common activities such as package and patch installations, it can transparently synchronize files to cluster nodes without any downtime or re-installation requirement. Updates are properly reflected in the Platform Cluster Manager repository so that changes will be retained the next time nodes are provisioned. For updates that do require a reboot, the web interface makes this apparent to cluster administrators, allowing them to defer the re-installation of nodes according to their own schedule. Perform risk-free, trial software installations Platform Cluster Manager employs a unique and flexible approach to software management supporting multiple repositories that in turn contain different operating systems and software components. Performing software upgrades becomes simple and risk-free. Administrators can snapshot their known good repository, and perform upgrades against a repository copy. If something goes wrong after deploying the upgraded software, node group definitions can simply be linked to the previous repository version, allowing the cluster to be quickly restored to its original state. This unique ability of Platform Cluster Manager, to create restore points for clusters, and the ability to maintain libraries of installable software components, is ideal for agencies or labs that may need to rapidly deploy earlier software stacks owing to regulatory requirements, or reliably demonstrate the repeatability of prior calculations. Flexible provisioning While Platform Cluster Manager, by default, employs the same provisioning approach as other cluster managers, it also offers enhanced capabilities. In addition to performing simple package-based installations, Platform Cluster Manager supports imagebased installs allowing cluster nodes to be rapidly cloned and even provides support for diskless nodes. Multiple operating systems and versions can be deployed concurrently to the same cluster, providing cluster administrators with the flexibility to support special requests from their user communities and accommodate unanticipated changes. The most advanced HPC tools Since performance is the whole reason for HPC clusters, Platform Cluster Manager provides the most advanced suite of HPC infrastructure components including libraries pre-optimized for vendor hardware and network components. Standard benchmark tests are included to ensure your cluster will deliver the best performance possible out-of-the-box. Platform Cluster Manager includes a range of industry standard, pre-tuned MPI implementations making it easy to get your parallel applications up and running quickly. For organizations requiring enhanced management of parallel workloads, Platform LSF, available separately, can be easily loaded as a kit and deployed to cluster nodes automatically. Exploit the power of GPUs in your HPC environment Available as a free download for Platform Cluster Manager, the NVIDIA CUDA Kit allows customers to easily harness the tremendous floating-point capability of modern graphics processors (GPUs) for general purpose computations. The NVIDIA kit automatically deploys a full CUDA 2.2 environment including all the drivers, tools, SDKs and debuggers needed to deliver breakthrough performance at low cost for a variety of widely used HPC algorithms optimized for NVIDIA GPUs.

8 Platform Cluster Manager Components Everything you need is here Vendor optimized Unlike other cluster deployment tools that may leave you scrambling for optimized drivers or specialty software components, Platform Cluster Manager is complete. It provides all the required software components so that open clusters are fully integrated out of the box. Platform Cluster Manager also includes a number of vendor specific enhancements including: The Intel Cluster Checker available on select configurations only, the Intel Cluster Checker is used to ensure Intel Cluster Ready (ICR) compliance. Vendor specific MPI run-time environments Enhanced InfiniBand support and optimized MPI libraries for selected switch manufacturers Enhanced device support for Mellanox, Myrinet, Qlogic, Vendor kits with enhanced drivers, tools and utilities specific to particular hardware vendors Fully supported by Platform Platform Cluster Manager is the only open Linux cluster solution fully supported by Platform Computing, the world leader in HPC management solutions. Unlike competing cluster toolkits which are little more than thrown together collections of open-source software, Platform Cluster Manager is the result of a close and long-standing engineering collaboration with our system partners. With a focus on quality and platform certification, Platform Computing and our systems partners jointly provide outstanding support on a 7 x 24 basis. When dealing with Platform, you re dealing with the organization that actually develops and builds the key software components comprising the cluster, ensuring top notch support and speedy problem resolution. Platform Cluster Manager includes everything you need to have your HPC cluster up and running quickly. For ease of installation and management, software is packaged as easy to deploy, pre-built kits. Of course, the Platform Cluster Manager cluster remains fully open, and administrators can still easily install their own software, using more traditional approaches if they choose. Kits are easy to build and sites may choose to package their own software as kits, or take advantage of a growing library of contributed kits from Platform Computing and others available at Among the pre-integrated kits included in Platform Cluster Manager are: Base Contains the tools and applications for managing the Platform Cluster Manager cluster Web GUI A complete web-based management interface and application-centric portal Cacti* An open source reporting tool used to collect and graph various node metrics Ganglia A powerful open-source resource monitoring solution HPC A comprehensive collection of tools, libraries and utilities specific to HPC clusters Platform Lava A powerful, open-source workload scheduler 100% compatible with Platform LSF Nagios A powerful open source host, service and network monitoring facility NTOP A tool to monitor network bandwidth and analyze traffic JAVA JRE Java 2 Platform Standard Edition Runtime Environment OFED A collection of drivers and libraries for InfiniBand * Note: Cacti is an open source rrdtool-based graphing solution. Copyright The Cacti Group For more information see

9 Built for Growth Most clusters grow not always in terms of nodes, but almost certainly in terms of applications, users and complexity as clusters are augmented with new hardware, network and storage devices. Platform Cluster Manager ensures that your cluster is future proof not only can you easily grow and scale, but you can support different operating systems and provisioning approaches as your needs evolve. With a unique driver patching capability, your cluster can even accommodate the addition of hardware not yet invented, avoiding the need to completely re-install or upgrade the cluster to support new devices. Should your requirements outgrow the open source components included, a range of optional commercial packages are available that seamlessly integrate with the Platform Cluster Manager cluster providing complete investment protection. Supported Operating Systems Platform Cluster Manager provides support for popular Linux environments including Red Hat Enterprise Linux, SUSE Linux and CentOS. The latest set of supported hardware platforms and operating environments is available at FOR MORE INFORMATION, CONTACT PLATFORM SALES SUPPORT: Tel: partnersales@platform.com Platform Computing is the leader in grid and cloud computing software that dynamically connects IT resources to workload demand according to business policies. Over 2,000 of the world s largest organizations rely on our solutions to improve IT productivity and reduce data center costs. Platform has strategic relationships with Cray, Dell TM, HP, IBM, Intel, Microsoft, Red Hat, and SAS. Building on 16 years of market leadership, Platform continues to help data centers be more efficient, responsive and dynamic. Visit World Headquarters Platform Computing Inc th Avenue Markham, Ontario L3R 3T7 Canada Tel: Fax: Toll-free tel: info@platform.com Sales - Headquarters Toll-free tel: Tel: North America New York: San Jose: Detroit: Europe Basingstoke: +44 (0) London: +44 (0) Paris: +33 (0) Düsseldorf: Munich: info-europe@platform.com Asia-Pacific Beijing: Xi an: asia@platform.com Tokyo: +81(0) info-japan@platform.com Singapore: wliaw@platform.com Copyright 2009 Platform Computing Corporation. The symbols and T designate trademarks of Platform Computing Corporation or identified third parties. All other logos and product names are the trademarks of their respective owners, errors and omissions excepted. Printed in Canada. Platform and Platform Computing refer to Platform Computing Corporation and each of its subsidiaries