The Infiniflow Service Fabric provides a distributed service oriented runtime that makes it easy to deploy, add resilience, scale, modify and manage your applications.

Your applications can be deployed directly on the Service Fabric allowing them to immediately take advantage of these benefits. With transparent support for applications developed in Spring and other POJO frameworks, Infiniflow removes the complexity associated with deploying and running applications across a distributed environment - allowing you to Develop Local, Deploy Global .

The Service Fabric is inherently resilient, scalable, flexible and adaptive - a platform that you and your business can depend upon.

Key features and benefits include:

A commercial time-limited evaluation release of Infiniflow is available to download today.

The base installation of Infiniflow, at less than 150kb per compute resource, offers an adaptive lightweight alternative to the traditional bloated and brittle application server and middleware deployment. The runtime real estate for any computer running Infiniflow depends upon the nature of the task you are asking it do, but Infiniflow's component-based architecture means that each compute resource only needs to install the code to deliver the functionality required at that moment - no more, no less.

Infiniflow's sophisticated Provisioner Service ensures that components that have specific minimum memory or CPU requirements only get deployed on those compute resources that can meet or exceed these criteria. Infiniflow also automatically uninstalls all unused services once the job is completed, and resources are returned to a clean state, ready to be automatically configured for the next requirement.

• Peaberry (formerly Guice-OSGi)
• iPOJO
• Qi4j

Infiniflow is operating system and hardware agnostic, and is complimentary to server virtualization strategies as it functions identically on physical or virtual machines.

A lightweight agent, the Atlas Agent (so named because it supports the rest of the world!), requires a one time install on every operating system instance. Once the Atlas Agent is installed, all software deployment and resource orchestration is fully automated - making it quick, easy and very cost effective to implement, change and manage the Service Fabric in a production environment.

Infiniflow is a Model-Driven architecture, where the components of a required runtime System, including business logic and required Infrastructure Services, are described in a SCA-compliant SCDL ( Service Component Definition Language) document called the Infiniflow System document. A System document may be thought of as an architectural blueprint or a recipe, describing how to assemble a runtime System.

Each System document lists the names and quantity of OSGi components required, together with the component version. It also defines how the components must be wired together to deliver the required functionality. The document can also specify absolute or preferred capabilities that the compute resource must provide in order to run a specific component. Each System document describes a model that equates to the Target State of the desired running System.

Given a System document, the Infiniflow Service Fabric will instantiate the prescribed functional elements, and then ensure the runtime entity continues to match the Target State described in the System document. Starting, pausing, modifying, enhancing or stopping the System is achieved by simply interacting with the System document. To modify the runtime system, simply change the System document and Infiniflow will do the rest.

The Infiniflow Service Fabric itself consists of over one hundred OSGi components, automatically deployed and assembled by activating the supplied Infiniflow Infrastructure Servce System documents. Systems can be made up of just a few components or 10's or 100's of components, the number determined by the desired granularity of each System. More on Model-Driven architectures >

Infiniflow offers a comprehensive Roles & Authentication Framework to control access to the System documents. For example, an analyst may be entitled to start, pause and stop a group of Systems within their business function, whereas a developer could be restricted to creating and testing a System but not running it within the production environment.

All interactions with each System document are logged and stored, and as applications can only be deployed, run, paused, modified or stopped by interacting with the System document, this creates a complete historical audit of what changes were made, when and by who. So with Infiniflow, the data to support Change Control and Audit is automatically collected as a result of the model-driven approach rather than as an after-thought. An optional Operational Service provides a GUI-based tool for accessing, collating and reviewing this data.

In a modern data center, where planned and unplanned change is the norm, treating provisioning as a one time activity makes the whole environment brittle. With Automated Continuous Provisioning, Infiniflow is able to deal with a fluid environment where servers come and go and components need to be modified and upgraded without downtime.

As part of its model-driven architecture, Infiniflow deploys a resilient Provisioner Service which continuously monitors every deployed Infiniflow Infrastructure Service and System, and compares the Target State defined in the System document with the current runtime state. If it detects a discrepancy, the Provisioner Service automatically makes the necessary changes to the runtime without the need for further human intervention. More on Provisioning >

Leasing is the mechanism used to detect failure in an Infiniflow Service Fabric and cause the Provisioner Service to deal with the problem.

When a component is deployed, a negotiation with the corresponding compute resource takes place, and the expected duration of the arrangement is defined in a Lease, which is then registered with the Service Fabric's Directory Service. Should a resource be removed or disappear, the lease will time out causing the Provisioner Service to compare the Target State in the System document with the runtime state to identify what components are now missing. The negotiation process will then be reactivated to automatically deploy replacement components until the Target State is achieved.

It should be noted that Systems will often be able to function, albeit at a sub-optimal level, without achieving Target State. In this scenario the Provisioner Service will keep trying to find suitable available compute resources to run the under-provisioned components.

With the ability to dynamically deploy components in a distributed environment, it is essential that Infiniflow provides an automated and elegant approach to removing components that are no longer required.

One of the key benefits of OSGi technology is comprehensive lifecycle management that makes it easy to install, start, stop, update and uninstall components. This enables 'zero' down-time patching/upgrading and the ability to add functionality on-demand, but it is limited to a single JVM.

Infiniflow extends this capability across a distributed environment - the Service Fabric - and enhances dependency management by providing an elegant distributed garbage collection and clean up capability.

The Infiniflow Service Fabric offers a highly resilient architecture with no single points of failure or special frames of reference. The self-healing capabilities exhibited by the Service Fabric are derived from the principles of Complex Adaptive System (CAS) design, and Infiniflow leverages Recovery Oriented Computing to, minimize recovery time rather than trying to stop things from failing.

Infiniflow maintains service availability by using sophisticated replication and group behaviour patterns to ensure that there are always sufficient Infrastructure Services available within a Service Fabric. Systems automatically benefit from the resilient Service Fabric, and Infniflow's Automated Continuous Provisioning ensures that Systems will always be running at the optimum level achievable at any point in time. More on Resilience >

Infiniflow is inherently scalable and has been designed to ensure that scaling up and down is simple and automated, and can be achieved dynamically if required. Infiniflow supports coarse and fine grained scalability for unrivalled flexibility and resource efficiency.

The conventional coarse grained approach where the entire application is replicated, is often effective but is far from efficient, as it means memory and compute cycles are wasted running multiple copies of the whole infrastructure and application stack. Fine grained scaling is a new approach where multiple instances of individual application components are deployed to remove a bottleneck. This approach is much more efficient and resource friendly, as all of the resource capabilities can be concentrated on the specific task that needs improving, but it demands a composite application approach and an innovative runtime, such as Infiniflow, to realize the full benefits.

Should application design prohibit fine grained scaling, Infiniflow's model-driven architecture provides significant benefits for conventional coarse grained scaling. Infiniflow makes it easy to deploy and manage many instances of an application, and automatically returns the hosting compute resources back to a clean state when an application instance is no longer required. More on Scalability >

The Fabric Administrator is an intuitive management GUI that is included free of charge with Infiniflow. This provides the ability to:

• load and unload SCA System documents and their associated OSGi components, to and from the Service Fabric's resilient CDS repository,
• start, stop and modify runnable Systems,
• manage and monitor the members of the compute resource population which together comprise the Service Fabric.

The Fabric Administrator fully leverages the Service Fabric's Roles & Authentication framework, allowing privileges to be assigned on an individual user or System basis.