Project Caroline Overview

Project Caroline is developing a horizontally scalable platform for the development and deployment of Internet services. The initial design center is to make the platform available as a utility, where a pool of virtualized resources is shared between many customers each with one or more services. Through the utility model services can dynamically flex their usage of the platform’s distributed resources up and down, matching usage to observed load. The horizontal scalability of the platform allows for the efficient delivery of resources and supports a pay-for-use (verses pay-for-capacity) billing model. Customers of the utility are isolated from each other and mechanisms are provided for the isolation of services.

The primary resource provided by Project Caroline is a set of horizontally scaled machines for running customer processes. Customers specify for each process the program to run, what networked file systems it should have access to, and IP addresses it should be reachable at. The platform takes care of the details of finding a machine for the process to run on, configuring the machine, network, and Internet connectivity. Operating system-level virtualization is used to isolate processes sharing the same physical machine while keeping per-process overhead low. Customer programs are expressed in languages like Java byte code, perl, and python that provide OS and instruction set independence. Other resources include IP sub-nets, network file systems, databases, external IP addresses, L4 and L7 load balancers, and DNS bindings. Applications can allocate, configure, and release these resources using the platform API. Through the platform API, applications can acquire and release resources in seconds.

In addition to the platform API, various tools and components have been layered on top of the platform API, including: cash, a Ruby-derived interactive shell; a standalone GUI and a NetBeans plug-in for direct manipulation of Project Caroline resources; a set of (J)Ruby centric tools for using Project Caroline; Apache Ant tasks for easily automating the creation and management of Project Caroline resources; and macro-components such as the Project Caroline Web Server (based on Project GlassFish v3) which automates setup and management of a horizontally-scaled web tier.

• Platform API JavaDoc
• Platform API Code Example
• Running Apache Tomcat On Project Caroline using Apache Ant
• Overview of Running Code On Project Caroline