FASTER

FASTER
Name	FASTER
Type	Technology/Protocol
Introduced	2010s
Developer	Consortium of industry and academic partners

FASTER

FASTER is a technology platform and protocol suite developed to accelerate high-throughput data transfer, real-time processing, and distributed coordination across heterogeneous networks. It integrates advances from hardware manufacturers, research universities, standards bodies, and industry consortia to provide low-latency, high-bandwidth solutions for scientific computing, multimedia distribution, and enterprise data pipelines. The design emphasizes modularity to interoperate with established systems from major vendors, research laboratories, and international projects.

Overview

FASTER combines ideas from Intel Corporation, NVIDIA, Cisco Systems, Amazon Web Services, Google Cloud Platform, Microsoft Azure, European Organization for Nuclear Research, National Aeronautics and Space Administration, Massachusetts Institute of Technology, Stanford University, ETH Zurich, University of California, Berkeley, Carnegie Mellon University, Lawrence Berkeley National Laboratory, Fermi National Accelerator Laboratory, Oak Ridge National Laboratory, European Space Agency, Max Planck Society, IBM, Oracle Corporation, ARM Holdings and others to bridge physical-layer optimizations with application-layer orchestration. It is used to coordinate transfers among storage systems like Ceph, Lustre, Amazon S3, and content-delivery infrastructures such as Akamai Technologies and Cloudflare. The platform targets workloads similar to those in projects like Large Hadron Collider, Square Kilometre Array, Human Genome Project, and media workflows at studios like Warner Bros., Netflix, Walt Disney Studios.

History and Development

Early research that influenced FASTER came from collaborations among labs at MIT Lincoln Laboratory, University of Illinois Urbana-Champaign, University of Cambridge, University of Oxford, Tsinghua University, Peking University, University of Tokyo, and industry research groups at Bell Labs, Xerox PARC, and Microsoft Research. Funding and pilot deployments involved agencies such as the National Science Foundation (US), European Commission, Japan Science and Technology Agency, Defense Advanced Research Projects Agency, and private foundations. Prototype implementations were demonstrated at conferences like SIGCOMM, Supercomputing Conference, USENIX, and International Conference on Data Engineering, with benchmark collaborations including OpenStack Foundation and Apache Software Foundation projects. Over successive revisions, standards contributions were made to bodies like IETF and IEEE Standards Association while interoperability tests ran at venues including Interop.

Technical Design and Specifications

FASTER's architecture layers hardware acceleration, transport optimization, and application-aware scheduling. It leverages accelerators from NVIDIA and Intel Xeon Phi families, programmable switches from Barefoot Networks/Intel Tofino, and RDMA-capable fabrics such as InfiniBand and RoCE. Control-plane integration uses agents compatible with Kubernetes, OpenStack, Ansible, and orchestration from platforms like HashiCorp Terraform. Protocol elements draw on research in QUIC, TCP, UDP extensions, and multicast techniques seen in PIM; security uses standards from TLS, IPsec, and OAuth 2.0. Storage interaction supports file systems and object stores including GlusterFS, HDFS, and Azure Blob Storage. Telemetry and observability plug-ins integrate with Prometheus, Grafana, ELK Stack, and tracing systems like OpenTelemetry.

Applications and Use Cases

FASTER is applied in scientific instruments and facilities such as Large Hadron Collider, Square Kilometre Array, James Webb Space Telescope, and ITER for streaming sensor data to analysis clusters. Media and entertainment workflows at Netflix, BBC, HBO, and Universal Pictures use it for live event distribution and studio postproduction. Cloud providers Amazon Web Services, Microsoft Azure, and Google Cloud Platform use similar patterns for cross-region replication, disaster recovery with partners like VMware, and edge services with companies such as Fastly. Enterprise analytics use cases include financial trading platforms at firms like Goldman Sachs and Citigroup and genomics pipelines at institutions like Broad Institute and Wellcome Sanger Institute.

Performance and Evaluation

Evaluations of FASTER implementations report substantial improvements in throughput and latency when compared to traditional TCP/IP stacks and commodity routing. Benchmarks performed at Supercomputing Conference and by teams from Lawrence Berkeley National Laboratory and Fermi National Accelerator Laboratory showed reductions in end-to-end latency for 4K/8K video streams used by NHK and higher aggregate transfer rates for large datasets comparable to results from Globus transfers. Performance tuning involves NIC offloads from Mellanox Technologies and switch pipeline optimizations from Barefoot Networks; results are typically validated using suites influenced by SPEC and TPC benchmarks and academic testbeds like PlanetLab and GENI.

Safety, Privacy, and Ethical Considerations

FASTER implementations must comply with regulations and guidelines from authorities such as European Commission directives, General Data Protection Regulation, and sectoral standards set by organizations like IEEE, NIST, and ISO. Privacy for human-subject data processed in collaborations with institutions like Broad Institute and Wellcome Sanger Institute is managed through data governance models influenced by HIPAA and national research ethics boards. Security reviews reference practices from CISA and incident response playbooks used by CERT Coordination Center. Ethical considerations include equitable access debated in forums such as United Nations and digital inclusion efforts promoted by World Bank and ITU.

Adoption and Impact

Adoption spans national laboratories, cloud operators, broadcasters, and research consortia. The ecosystem effect influences procurement at agencies like NASA, DOE, and European Space Agency, and informs standards work at IETF and IEEE Standards Association. Academic curricula at institutions such as MIT, Stanford University, Carnegie Mellon University, and ETH Zurich incorporate related topics into courses and research projects. The convergence of industry players like IBM, Cisco Systems, Amazon Web Services, Google, and Microsoft around similar optimization goals has accelerated practical deployments and collaboration across national and commercial boundaries.

Category:Computer networking