liveBook · Manning

Load balancing

As explained in Event Processing in Action

The distribution or adjustment of work among available computing resources to manage workload.

In event-processing systems, load balancing concerns reevaluating how work is partitioned among computing resources as conditions change. Static partitioning assumptions can become outdated when resources or event loads shift. section 10.4.1

Two responses to workload and resource constraints are described:

  • Reevaluate the partitioning scheme when resources are added or temporarily unavailable, or when event distributions and loads change. General guidance on when and how to rebalance remains unresolved. section 10.4.1
  • Discard events when available resources cannot process them all. This approach, called load shedding, produces approximate event processing. section 10.4.1

How Event Processing in Action covers it

How deep it goes
Distinguishes partitioning reevaluation from load shedding but leaves general decisions about when and how to rebalance unresolved.
Examples
Uses resource changes, shifting event loads, and event discarding under insufficient capacity as cases.
What you can do after
Understand why changing resources and event loads can prompt repartitioning, and recognize event discarding as load shedding.
How it connects
Connects load balancing with partitioning, stratification, parallel processing, and load shedding.

Also explained in

Compare how 15 books explain load balancing →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Event Processing in Action. The section links point to the parts of the book it draws on.