Advanced Networking

Advancing state of the art high-performance computing, communications and data analytics.

The Advanced Networking group at Pittsburgh Supercomputing Center conducts research on network performance and analysis in support of high-performance computing applications. They also develop software to support heterogeneous distributed supercomputing applications and to implement high-speed interfaces to archival and mass storage systems.

Work done by the Advanced Networking group includes TCP implementations, TCP tuning, and software to monitor and improve network performance.

Research Projects

3ROX –  The Three Rivers Optical Exchange (3ROX) is a regional network aggregation point, also called a GigaPoP, providing high speed commodity and research network access to sites in western and central Pennsylvania and West Virginia.

Web10G – Web10G promises innovations in performance, stability, optimization, and network diagnostics  to anyone who wants to fully exploit modern high bandwidth connections.   It provides users with have real-time statistics about their network connections, and allows them to fine-tune some characteristics of the connection to improve performance. Join the Web10G team on our mailing list, view the archives  or visit Web10G online!

KINBER and PennREN – The Keystone Initiative for Network Based Education and Research (KINBER) serves as coordinator for the construction and management of a Pennsylvania-wide fiber optic network, PennREN (Pennsylvania Research and Education Network).  PennREN will be accessible to educational, research, health care, and economic development partners seeking to aggregate services for their members and subscribers at affordable cost.

  

Networking Tools, software, and information

TestRig 2.0 – A bootable Linux ISO used to automatically and conveniently generate network performance data for analysis by PSC engineers.

High Performance Enabled SSH/SCP – HPN-SSH Maximize the performance of secure copying and connections between hosts on a high-performance network.

Enabling High Performance Data Transfers

 

TCP performance tuning

  • TCP Performance Tuning
    • Presentation by Jamshid Mahdavi at the High Performance Data Networking conference at Northwestern University, 4/97
    • Presentation by Jeffrey Semke at the vBNS Techs workshop at Carnegie Mellon University, 5/98
  • Poll-Switch – SNMP poll a layer-2 device – Retrieving host and port information from a layer-2 switch running an SNMP agent.
  • Mac-Scan – Scan hosts on a VLAN or network for vulnerabilities
  • IP-Utils – Enable data communication via a simple API for various IP-based networking protocols

Networking Research Papers

 

Previous work

  • DANCES – The DANCES project (Developing Applications with Networking Capabilities via End-to-End SDN) is  developing tools to allow the scheduling of network resources much as compute resources are scheduled now, with the aim of reducing data transfer bottlenecks, a serious hindrance to scientifc research.
  • NPAD/pathdiag Automatic diagnostic server for troubleshooting end-systems and last-mile network problems.
  • SAMI (NIMI) Creating a Secure and Accountable Measurement Infrastructure
  • Web 100 Created to address automatic TCP buffer tuning.

Learn more about  the ability to maximize the performance of secure copying and connections between hosts on a high-performance network with HPN-SSH.

Highlighting HPN-SSH

Much of modern research requires Big Data transfers between collaborators at opposite ends of the country. The SSH (Secure SHell) protocol has allowed scientists to safely trade encrypted data over insecure networks, removing the trouble and expense of securing every component in a data transfer. But it wasn’t designed for today’s data loads and has many bottlenecks. In 2004, PSC staff published HPN-SSH, a set of patches to remove those bottlenecks. But a lack of dedicated funding has forced them to do only maintenance work on this now-popular package of networking tools. A new NSF grant will now allow PSC to upgrade HPN-SSN, adding features like additional acceleration, allowing data transfers to re-start at the point of failure instead of at the beginning, and telemetry readouts to aid in troubleshooting and performance analysis.