A. PaNDA (Platform for Network Data Analytics)!

 

With Cisco Live Berlin approaching fast we'd like to invite you to come and see PaNDA!


PaNDA is for a simple, scalable, open big data / analytics platform analytics platform supporting operational intelligence (e.g. OSS) and business intelligence applications (e.g. BSS) applications for networks and network services.

 

We built PaNDA because - in trying to work out how OSS and BSS evolved and how we could apply Big Data technologies to the data coming from networks and services - we realised that many of the functions performed by OSS/BSS today are really best addressed as big data problems, i.e. they can be addressed by performing an arbitrary function against the entire OSS data set, e.g:

  • Fault management = f(event data)
  • Performance management = f(metric data)
  • Billing mediation = f(event data, metric data)
  • Capacity management = f(metric data)
  • Security analytics = f(metric data, route data)

 

The PaNDA architecture (shown below) is based on simple, scalable, open data platform, which leverages best current practise for Big Data infrastructures:

  • Scale-out architecture with support for horizontal scale in all core components
  • Very highly available core platform
  • Low and predictable latency
  • Pub/sub system with very high scale message distribution between data aggregation and analysis functions
    • Functions can be both consumers and publishers
    • Open APIs to make it simple to publish and consume
  • Immutable dataset with minimal filtering/processing on ingress
  • Support a combination of high-volume, low-latency and real-time processing applications

 

Screen Shot 2016-02-11 at 10.29.18.png

 

Such a system architecture for OSS analytics has the following benefits:

  • An open system architecture
  • Allows any analysis application to mine any data source, exposing the full value of the OSS dataset
  • Extensible – add new OSS analysis functions quickly and seamlessly with minimum of development cost
  • Removes duplicate polling/collection – collect data once, use many times
  • Remove the need for cross application integration – applications can be both publishers and consumers
  • Ability to handle heterogeneous signals/data – structured and unstructured, e.g. logs, events, metrics/time-series data, and network telemetry,
  • Leverage rapid innovation in Big Data analytics space
  • Platform is extensible beyond OSS/BSS and forms a generalised platform for network analytics.

 

Hence, if you are trying to understand how to apply big data technologies to the data coming from your networks and services, or how the OSS space will be impacted by Big Data, please drop by the DevNet Zone and see the PaNDA demos, meet Team PaNDA or even try out the code.

 

See you there!

 

Team PaNDA.

 

PaNDA will be featured at the Cisco Live 2016 Berlin | February 15-19 | Networking Event. Come to see the demo in the DevNet zone, play with some learning labs and attend the session with John Evans / Nick Hall / Trev Smith, more information here: Content Catalog | Cisco Live 2016 Berlin | February 15-19