Key Takeaways from Our Oct 13, 2016 Podcast: "Putting Machine Data to Work"

Author: 
Vijay Vasudevan
Monday, October 17, 2016

Putting Machine Data to Work

 

 

 

 

 

 

 

 

This is a tiny recap on the webinar while focusing on some of the most pertinent questions asked by the participants. In case you missed the webinar, you’re likely to find some useful insights by just going over the Q&A session we’ve tried to summarize here.

Bret, Ryan, and Puneet collectively traversed the entire chain of churning machine logs to snappy dashboards. They underscored the ecosystem’s power to metamorphose complex industrial machine data like those generated in the converged infrastructure space. While Puneet and Ryan focused on showcasing the underlying technology stack, Brett on the other hand discussed how Springpath has branded the PTC-Glassbeam IoT solution and deployed it over the counter for CISCO.

Armed with a unique solution, Springpath is now pulling in tens of thousands of CISCO nodes onto its hyperconverged infrastructure. Ops personnel at CISCO are getting IoT-machine data based insights that are not being experienced by any other product manufacturer in the Silicon Valley.

Don’t take our word for it. Hear the recorded podcast here.

Now, let’s go over some of the Q&As.

Is the PTC-Glassbeam solution best implemented in a gateway-based infrastructure as opposed to aggregating directly from center-center platforms?

  • It’s a cloud-based solution that runs on the Amazon infrastructure or Dimension Data
  • Clients are expected to send data to the cloud and the aggregation happens there
  • Glassbeam Edge is a light-weight deployment of the cloud-based solution
  • GB Edge strips all of the storage, back-end heavy weight lifting and allows you to perform ETL, Parsing right at the device point
  • Fits perfectly with PTC’s ThingWorx Edge platform
  • Roadmap for Glassbeam Edge GA is in the works

In the ThingWorx Analytics solution is there a machine learning component? What is it based on?

  • A big component of the solution is deep machine learning algorithms
  • Used in understanding and predicting machine failure
  • Algorithms use super vector machines, artificial neural networks, linear logistic regression, decision trees, and so on
  • Layer above the models determine what to use under what scenarios
  • Aim to reduce human interaction to determine what techniques to use in varied situations

We had other questions on whether a data scientist is required as part of this solution architecture? Are custom solutions available? Is there a latency involved in processing the huge velocity of data?

Should you be interested in knowing what panelist’s response were to those questions or just want catch up on the rest of the webinar, please click here.