dCache is a distributed storage system proven to scale to hundreds of Petabytes. Originally conceived as a disk cache (hence the name) in front of a tertiary storage to provide efficient data access for data intensive scientific experiments in the field of High Energy Physics (HEP) it has evolved into highly scalable general-purpose open source storage solution.
dCache is a joint effort between the Deutsches Elektronen-Synchrotron (DESY) in Hamburg, Nordic Data Grid Facility (NeIC, NDGF) based in Copenhagen, the Fermi National Accelerator Laboratory near Chicago with significant distributions and support from the University of California, San Diego, INFN, Bari as well as Rutherford Appleton Laboratory, UK and CERN in Geneva.
A dCache instance will generally consist of many storage (or “pool”) nodes. On those nodes, normal Linux filesystems (btrfs, ext4, XFS, ZFS) are used to store data.
Alternatively, dCache pools can use storage space provided by a Ceph object storage system.
In addition to those possibilities, dCache can use its hierarchical storage management capabilities to transparently use storage systems with different characteristics (like tape libraries for lower-cost, but higher-latency storage). Built-in mechanisms can be used to increase performance and balance loads, increase resilience and availability. dCache also supplies advanced control systems to manage data as well as data flows.
dCache supports the following I/O (and data management) protocols:
- FTP (including GridFTP)
- HTTP (and WEBDAV)
- NFS (parallel NFSv4)
dCache supports X.509 certificate based authentication through the Grid Security Infrastructure used as well as username/password authentication and LDAP. For some workloads, users can also authenticate using macaroons or OpenID Connect.
dCache provides fine-grained POSIX and NFS-style access control list (ACLs) based file/directory authorization.
Resilience and high availability can be configured by enabling multiple file replicas and flexible variety of replica placement policies.
Easy migration of data via the migration module.
The “billing” system provides a powerful cost calculation system that allows to control the data flow (reading and writing from/to data servers, between data servers and also between data servers and tape).
Load balancing and performance tuning by hot pool replication (via cost calculation and replicas created by pool-to-pool-transfers).
Space management and support for space tokens.
Detailed logging and debugging as well as accounting and statistics.
XML information provider with detailed live information about the system.
Scriptable adminstration interface with a terminal-based front-end.
Web-interface with live information of the most important information.
Automatic checksumming for data integrity.
For a minimal test installation:
Hardware: contemporary CPU, 1 GiB of RAM , 500 MiB free hard disk space
Software: OpenJDK 8 or later, Postgres SQL Server (9.6 or later)
For high performance production scenarios, the hardware requirements greatly differ, which makes it impossible to provide such parameters here. However, if you wish to setup a dCache-based storage system, just let us know and we will help you with your system specifications. Just contact us: email@example.com.
This book shall introduce you to dCache and provide you with the details of the installation. It describes the configuration and customization of dCache as well as the usage of several protocols that dCache supports. Additionally, it provides cookbooks for standard tasks.
Here is an overview part by part:
Part 1, Getting started: This part introduces you to the cell and domain concepts in dCache. It provides a detailed description of installing, the basic configuration, and upgrading dCache.
Part 2, Configuration of dCache: Within this part the configuration of several additional features of dCache is described. They are not necessary to run dCache but will be needed by some users depending on their requirements.
Part 3, Cookbook: This part comprises guides for specific tasks a system administrator might want to perform.