How Milvus Balances Query Load Across Nodes

In previous blog articles, we have successively introduced the deletion, bitset, and compaction functions in Milvus 2.0. To culminate this series, we would like to share the design behind load balance, a vital function in the distributed cluster of Milvus.

Usage

Milvus 2.0 supports automatic load balance by default. But you can still trigger load balance manually. Please note that only sealed segments can be transferred across query nodes.

Best Practices for TiDB Load Balancing

Load balancing distributes connections from applications to TiDB Server instances. This helps to distribute the load over multiple machines and, depending on the load balancing option, can automatically reroute connections if a TiDB instance becomes unavailable.

Load Balancing Types

There are many different ways to implement a load balancer. This section describes the most common types.

Mulesoft Dedicated Load Balancer Use Case

Introduction

We have recently implemented dedicated load balancers (DLB) into our integration landscape. I am sharing details about our use case.

Before having DLB, the VPC received all the traffic through the shared load balancer (SLB) or directly to mule internal workers through VPN. There were two limitations to this approach:

Top 10 Tips for Making the Spark + Alluxio Stack Blazing Fast

The Apache Spark + Alluxio stack is getting quite popular particularly for the unification of data access across S3 and HDFS. In addition, compute and storage are increasingly being separated causing larger latencies for queries. Alluxio is leveraged as compute-side virtual storage to improve performance. But to get the best performance, like any technology stack, you need to follow the best practices. This article provides the top 10 tips for performance tuning for real-world workloads when running Spark on Alluxio with data locality, giving the most bang for the buck.

A Note on Data Locality

High data locality can greatly improve the performance of Spark jobs. When data locality is achieved, Spark tasks can read in-Alluxio data from local Alluxio workers at memory speed (when ramdisk is configured) instead of transferring the data over the network. The first few tips are related to locality.