Open Cloud Conundrum, Open Cloud Consortium

One of the hot areas in 2008 was cloud computing, and 2009 looks likely to be a year that is equally occupied with the subject. But cloud computing represents something of a conundrum for the open source world.
##CONTINUE##
Much of it is built using free software infrastructure - naturally enough, since it scales well both in terms of performance and cost. But it's not clear from a legal viewpoint whether providing cloud computing services constitutes distributing software in the sense of traditional free software licences like the GNU GPL.

That represents something of a loophole - one that the GNU Affero GPL seeks to close. Another issue is to do with data portability, or lack of it. Just because the back-end software is open doesn't mean you can get your data out easily. And so we have the paradoxical situation where free software is being used to create data lock-in.

It's still early days for cloud computing, so maybe it's too much to expect complete answers to these problems. But at the very least we need to be thinking about how to go about coming up with those answers. Here's s omething that might help , the Open Cloud Consortium, which:

1. supports the development of standards for cloud computing and frameworks for interoperating between clouds;
2. supports the development of benchmarks for cloud computing;
3. supports open source software for cloud computing;
4. manages a testbed for cloud computing called the Open Cloud Testbed;
5. sponsors workshops and other events related to cloud computing.

In truth, there's not much there yet, but at least it seems to be making the right noises. Here's some of the open cloud computing software it lists:

Hadoop

Hadoop is the most wide deployed open source software for cloud computing. It is available from hadoop.apache.org/core.

Hadoop is deployed on all the nodes of the Open Cloud Testbed.

Thrift

Thrift is a software framework for scalable cross-language services development. It integrates a software stack with a code generation engine to build services that work efficiently between C++, Java, Python, PHP, and Ruby. Thrift was developed at Facebook, and is available as open source at [http://incubator.apache.org/thrift/].

Thrift is expected to be available shortly on the Open Cloud Testbed and will used to interoperate different cloud middleware and to simplify the development and deployment of cloud applications.

The last of the cloud computing apps on this page is slightly different, because it is being supported directly by the OCC; it's one that I've not come across before:

Sector

The Open Cloud Consortium (OCC) is supporting the development of Sector, an open source stack for cloud computing. Sector is available on all the nodes of the Open Cloud Testbed.

In benchmarks using Terasort, Sector is about twice as fast as Hadoop. A paper describing these results will be presented at the upcoming KDD 08 conference. The paper is also available from the Sector web site sector.sf.net.

Sector is designed to operate over wide area clouds. Sector is built using UDT, which is a network transport protocol designed for wide area high performance networks.

Sector is open source and available from Source Forge at the web site: sector.sf.net

Here is some more information about the software from its home page:

Sector/Sphere can be broadly compared to Google's GFS/MapReduce stack. However, Sector supports wide area deployment, while Sphere is more flexible because it supports user defined functions that can be either map, reduce, or others. Below is a summary of the key features of Sector/Sphere.

* Provides high performance distributed storage. Sector can also be installed over wide area networks, thus it can be used as a content distribution network.
* Provides persistent data storage. Sector automatically replicates data files in the systems.
* Provides simplified distributed application development support. Each processor in the system can process a data segment with any user defined function. Data location, movement, and fault tolerance is transparent to developers.
* Provides both data access tools and programming API in C++.

It's hard to tell whether the Open Cloud Consortium will become one of the key organisations driving the evolution of open source work in this area, but it's good that people are at least starting to address the issue.

-----------------------------
BY Glyn Moody
Source：COMPUTERWORLDUK