OpenSAF

OpenSAF (commonly styled SAF, the Service Availability Framework[1]) is an open-source service-orchestration system for automating computer application deployment, scaling, and management.

[3] OpenSAF is the most complete implementation of the SAF AIS specifications, providing a platform for automating deployment, scaling, and operations of application services across clusters of hosts.

[4] It works across a range of virtualization tools and runs services in a cluster, often integrating with JVM, Vagrant, and/or Docker runtimes.

Membership evolved to include Emerson Network Power, SUN Microsystems, ENEA, Wind River, Huawei, IP Infusion, Tail-f, Aricent, GoAhead Software, and Rancore Technologies.

[8] OpenSAF's development and design are heavily influenced by Mission critical system requirements, including Carrier Grade Linux, SAF, ATCA and Hardware Platform Interface.

[14] OpenSAF 3.0 released on June 17, 2009, included platform management, usability improvements, and Java API support.

OpenSAF defines a set of building blocks, collectively providing a mechanism to manage Service Availability (SA) of applications based on resource-capability models.

[21] OpenSAF is designed for loosely coupled systems with fast interconnections between nodes (i.e. using TIPC/TCP),[22] and extensible to meet different workloads; components communicate between themselves using any protocol.

This extensibility is provided in large part by the IMM API, used by internal components and core services.

The platform can exert control over compute and storage resources by defining as Objects, to be managed as (component service) instances and/or node constraints.

Ongoing research to address this gap,[25][26] needs to deliver ecosystem tools, to better support modeling and automation of carrier-grade and Cloud Native Computing Foundation use cases.

The SU is an instantiable unit of workload controlled by an AMF redundancy model, either active, standby, or failed state.

It can be used to guarantee the availability of a specified number of identical SU's based on selected configured redundancy model: N-Way, N-way-Active, 2N, N+M, or 'No-redundancy'.

[2][16] An OpenSAF Service Instance (SI) is a set of SU that work together, such as one tier of a multi-tier application.

Multi-instance SG (N-way-active, N-way, N+M) requires a stable IP address, DNS name, and load balancer to distribute the traffic of that IP address among active SU in that SG (even if failures cause the SU's to move from machine to machine).

One solution is a Network File System (NFS) shared storage, accessible to all payload nodes.

[30] Other technical solutions are possible - what is important is that Volumes (File Share, mount point) can be modeled in AMF.

Many network equipment providers have demonstrated their support for OpenSAF by joining the Foundation and/or contributing to the Open Source project.

OpenSAF is functionally complete but lacks the ecosystem of modeling tools available to other open-source solutions like Kubernetes and Docker Swarm.

Service Availability, Principles and Practice, Textbook
OpenSAF v4 Architecture
System Designers need better Modeling tools
OpenSAF Concepts