OpenAFS

OpenAFS
Name	OpenAFS
Developer	Hewlett-Packard Corporation; IBM; Transarc; Collaborations among Carnegie Mellon University researchers, University of Michigan contributors, Stanford University engineers
Initial release	2000 (as open source continuation)
Written in	C; C++
Operating system	Linux; Microsoft Windows; macOS; FreeBSD; NetBSD; OpenBSD
License	IBM Public License; Open Source Initiative-compatible

OpenAFS is a distributed networked filesystem implementation descended from the Andrew File System research project, providing a wide-area, location-transparent filesystem for heterogeneous environments. It builds on work by corporations and universities to offer cross-platform file sharing with integrated replication, caching, and access control, and is used in academia, enterprise, and research institutions. OpenAFS interoperates with diverse identity and security systems and supports large-scale deployments spanning thousands of clients and servers.

History

OpenAFS originated from the commercial and academic lineage of the Andrew File System developed at Carnegie Mellon University and commercialized by Transarc Corporation, later acquired by IBM. Following vendor changes and evolving open source trends exemplified by projects like Linux kernel development and the Apache Software Foundation, OpenAFS emerged to continue maintenance and enhancement under community governance. Major contributors include engineers from Hewlett-Packard, IBM, and researchers with ties to Stanford University and University of Michigan, paralleling collaborative efforts seen in projects such as BSD and X Window System. Over time OpenAFS integrated protocols and practices influenced by Kerberos, LDAP, and distributed projects like NFS and Coda file system.

Architecture

The architecture implements a client–server model with distinct roles: fileservers (volumes served by pservers), volume location services similar to concepts in distributed systems like DNS, and client cache managers that mirror caching techniques from Squid (software) and Akamai Technologies. Volumes are logical unit groupings allowing operations analogous to Logical Volume Manager management and snapshot strategies used in ZFS. Replication and backup workflows align with approaches from rsync and enterprise storage solutions by EMC Corporation and NetApp. Authentication and authorization are designed to interoperate with MIT Kerberos realms and identity directories such as OpenLDAP and Microsoft Active Directory, reflecting federated identity patterns used by InCommon and EduGAIN.

Implementation and Features

OpenAFS is implemented primarily in C and C++, with platform-specific kernel modules and userland components comparable to device driver projects in the Linux kernel and filesystems like ext4 and XFS. Key features include client-side caching, callback-based cache coherence inspired by research in distributed caching, volume management resembling operations in LVM and Btrfs, and tiered storage strategies used by vendors like IBM and Hewlett-Packard. The filesystem exposes POSIX-like semantics comparable to POSIX-compliant systems and integrates with backup tools such as Bacula (software) and snapshot utilities akin to Time Machine (macOS). OpenAFS supports cross-platform clients on Microsoft Windows via service and driver components, paralleling compatibility efforts seen in Samba (software) and CIFS integration.

Deployment and Administration

Administrators deploy OpenAFS in topologies that mirror large-scale directory and service configurations found in deployments of Apache HTTP Server and PostgreSQL clusters. Centralized cells and cells’ volume location databases require coordination similar to operations of DNS zones and Active Directory forests. Management tooling follows patterns from Nagios and Prometheus (software) for monitoring, and configuration management often uses systems like Ansible, Puppet (software), and Chef (software). Operations for backup, disaster recovery, and migration draw on procedures established by practitioners of VMware virtualization and enterprise storage migrations at organizations such as Facebook and Google research groups.

Security and Authentication

Security in OpenAFS hinges on integration with Kerberos ticketing as used in academic federations like InCommon and directory services such as OpenLDAP and Microsoft Active Directory. Access control lists and rights management parallel models used by POSIX permissions and enterprise identity governance from Okta and Ping Identity. Secure transport and host-based controls reflect practices promoted by National Institute of Standards and Technology publications and hardening guidance from CIS (Center for Internet Security). Auditing and logging interoperate with centralized logging systems like syslog and Splunk, akin to security operations at institutions such as NASA and MITRE.

Performance and Scalability

OpenAFS employs callback-based cache coherence and client-side caching comparable to techniques developed for Akamai Technologies and web caching architectures in Squid (software), enabling high performance across WANs and LANs. Scalability patterns follow distributed file system research exemplified by Google File System and Hadoop Distributed File System, with considerations for volume partitioning, replication, and load balancing similar to Ceph and GlusterFS. Performance tuning often references kernel TCP/IP stack optimizations in the Linux kernel and storage I/O strategies used by NetApp and EMC Corporation appliances. Large academic deployments at universities such as University of California, Berkeley and University of Cambridge illustrate real-world scaling practices.

Community and Licensing

The project is stewarded by an open-source community that parallels governance models used by the Apache Software Foundation and FreeBSD project, with contributions from corporations like IBM and Hewlett-Packard as well as universities such as Carnegie Mellon University and Stanford University. Licensing uses the IBM Public License, aligning with open-source distributions overseen by the Open Source Initiative, and collaboration channels mirror those of projects hosted on platforms similar to GitHub and GitLab. Community activities include mailing lists, conferences akin to USENIX symposia, and interoperability testing similar to IETF working group practices.

Category:Distributed file systems