Traditional vs. workload-aware auto-scaling: What suits big data?

Workload-Aware Auto-Scaling

Cover

When it comes to big data applications, you might not be able to rely on traditional auto-scaling methods.

Those original auto-scaling technologies often work just fine for standard applications, but they can be a poor fit for big data for several reasons, including:

  • Removing nodes while downscaling is hard
  • Downscaling algorithms need to pick nodes carefully
  • Same auto-scaling policies can’t be applied to all workloads
  • And more

A workload-aware auto-scaling technology might be the better fit, especially to orchestrate with Hadoop and Apache Spark. Download this resource for more information.

Vendor:
Qubole
Posted:
28 Feb 2018
Published:
28 Feb 2018
Format:
PDF
Length:
13 Page(s)
Type:
Resource
Language:
English
Already a Bitpipe member? Login here

Download this Resource!