Fremont, CA 94538
+1 (510) 982-1003
info@broadbridgenetworks.com

Technology

Created with Sketch.
GPU ACCELLERATED END-TO-END DATA PIPELINE

An all new data pipeline, optimised for AI and analytic workloads, reimagined down to the data frame

Read more
REAL TIME ANALYTICS WORKBENCH

Cross filtered analysis allowing a multi-dimensional exploration with unparalleled interactivity

Read more
GPU BASED HARDWARE ARCHITECTURE

Deployed on Nvidia’s DGX & EGX Server platforms, providing unparalleled performance at scale

Read more

A Complete Reinvention of the
Cyber Security Data Pipeline

Because Latency Matters

The advantages offered by modern GPU computing are fully realised with BroadBridge Networks’ entirely redesigned software stack, which incorporates breakthrough advances in network transport, disk IO, memory frameworks, and accellerated AI data science libraries

OPTIMISED TRANSPORT FRAMEWORK

A revoliutionary “on-the-wire” data representation that does not require deserialization on receipt.  Anticipating future requirements, we support parallel transfers, allowing data to be streamed to or from a cluster of servers simultaneously.

IO OPTIMIZATION STACK

Innovative IO optimization technologies open a direct data path between GPU memory and storage, avoiding the CPU altogether. This increases bandwidth, decreases latency, and decreases utilization load on the CPU and GPU.

GPU OPTIMIZED VISUALIZATION

We capitalise on developments in graphics pipeline systems for creating meaningful graphical, geospatial, and network representations of large analytic datasets in milliseconds.

CROSS FILTERED ANALYTIC PARADIGM

Cross filtered analysis provides a natural interface for deep data exploration, allowing a multi-dimensional exploration with unparalleled interactivity when drilling into data for outliers, trends, or anomalies.

Analyst Workbench

Real time streaming data workbench with customizable charting and plotting, supporting multiple concurrent analysts, providing both geospatial and network graph analysis & full cross-filtered graph analytics. 

GPU optimization provides memory bandwidth exceeding 20x system memory, allowing exploration of multi-billion-record datasets in milliseconds

ANALYST WORKBENCH

unparalleled interactivity
when drilling into data for outliers, trends, or anomalies.

Cross filtered analysis provides a natural interface for deep data exploration, allowing a multi-dimensional exploration with unparalleled interactivity.

With fully integrated AI based data enrichment and analysis tools, analysts can quickly isolate threats, pivot on any data attribute, and drill quickly into temporally relevant network logs.

Advanced analysts can push analytic result sets into Jupyter notebooks with the click of a button, opening up a new dimention in forensic analysis.

GPU Hardware Platform

GPUs provide the horsepower required by compute-intensive workloads, but their data consumption increases the demand for input/output (IO) bandwidth. BroadBridge integrates an all new end-to-end data pipeline with highly optimised networking, file systems, and storage to maximize IO performance for multi-GPU, multi-node accelerated systems.

EGX Platform for
Edge Computing

ENTERPRISE PERFORMANCE WITH UP TO 4 GPUs

For enterprise customers BroadBridge LegionTM is deployed on NVIDIA’s EGX A100 platform (Deploying on V100 based EGX through Q3 20).  EGX A100 features a highly optimized software stack for GPU accellerated NGC-Ready for Edge servers. The stack includes NVIDIA drivers, a Kubernetes plug-in, a container runtime, and containerized AI frameworks and applications, including NVIDIA TensorRT™, TensorRT Inference Server, and DeepStream.

LEARN MORE

NVIDIA DGX A100 Supercomputer

TELECOM SCALE ‘ALL NETWORK’ PERFORMANCE

To meet the demands of major telecom scale networks, the BroadBridge LegionTM cybersecurity platform is deployed on NVIDIA® DGX-A100™ systems, capable of delivering a massive 5 Petaflops of compute in just 6RU.  With DGX A100, multiple I/O streams, at line rates of 200GBit/sec, can be supported at scale.

Larger, continental scale networks, can be supported on Nvidia’s DGX POD and DGX Super POD with up to  5,242,880 CUDA cores and 655,360 Tensor cores.

LEARN MORE