Fluid: Data Orchestration for Kubernetes
Shriira Press
Stop data from bottlenecking your GPUs. Orchestrate and accelerate data-intensive Kubernetes workloads with Fluid — caching, prefetching, and data locality.
Welcome to Fluid: Data Orchestration for Kubernetes.
Fluid is a Kubernetes-native distributed dataset orchestration and acceleration engine for data-intensive applications (AI/ML and big data) — making data a first-class, cached, locality-aware citizen of Kubernetes. This free book teaches it from the ground up: the data problem in cloud-native compute and what Fluid is, the data-locality problem (decoupled storage, repeated fetching, idle GPUs), Fluid's architecture (Datasets, Runtimes, controllers), the Dataset abstraction, caching runtimes (Alluxio, JuiceFS, ThinRuntime), data acceleration (caching, prefetching, fast reads), data-aware scheduling (bringing compute to the data), AI/ML and big data use cases (training, serving, analytics), operating Fluid (managing caches, observability, autoscaling, consistency), and using Fluid in practice. Ten focused chapters with clear diagrams that demystify how to keep expensive compute fed — by caching data near compute, reusing it, and scheduling for locality — so data stops being the bottleneck.
This title is part of the ShriIra library and is free to read in full, right here — our small contribution to making world-class knowledge easy to reach.
A note on reading it: open the Contents menu at the top of the reader to jump between chapters, use the Aa menu to set a comfortable text size, theme (light, sepia, or night), and single- or two-page layout. Your place is saved automatically, so you can always pick up where you left off.
We hope it serves you well.
— Shriira Press