Google Compute Engine

Chapter 3. Storage: Persistent Disk

Compute Engine provides a rich array of storage mechanisms, which we’ll describe in detail in this and the next two chapters. Along the way, we’ll learn how to use each mechanism via our three main access methods: the Developers Console Web UI, the gcloud command-line tool, and the Compute Engine API.

Compute Engine Storage Options at a Glance

Compute Engine supports numerous storage options. Persistent disk (PD) is the primary mechanism available for block storage and is the method of choice for storing filesystem data. Virtual machines boot from persistent disks and also use PDs to store supplementary (i.e., nonboot) filesystems or raw data. This chapter describes persistent disks in depth.

Persistent disks are local storage devices. They are available from Compute Engine virtual machines only. The other storage options we’ll cover in this book are available both within a VM and outside of Compute Engine as well. Google Cloud Storage, which provides an Internet-wide, high capacity, highly scalable binary object store, is described in Chapter 4. Cloud SQL and Cloud Datastore (both described in Chapter 5), provide managed MySQL database instances, and a highly scalable global NoSQL database in the cloud, respectively.

Table 3-1 provides a summary of the storage mechanisms available to Compute Engine users, along with several key attributes of each.

Table 3-1. Overview of storage options

Option7	Data model	Managed	Storage scope