You want it Fast or Super Fast? – The IB, 10GbE , GigE Benchmark

Shay Hassidim — Sun, 16 Nov 2008 18:46:49 +0000

With the current global financial meltdown, the ability to effectively compete becomes essential. Faster data access and sharing are critical for business success. Speed is critical for beating the competition, translating into a need for better latency. With MW products this means the ability to push your data from the client end point into the clustered [...]

Big Cloud Week Coming Up

Geva Perry — Fri, 14 Nov 2008 22:23:45 +0000

Next week is shaping up to be a big week for cloud computing (again). MondayMicrosoft will be making an announcement in San Francisco about the launch of Online Services. I've been invited to attend that as well as a Blogger...

Private/Public Cloud

Nati Shalom — Fri, 14 Nov 2008 21:00:41 +0000

Most data centers of today run applications on dedicated machines. This is often referred to as static provisioning. In addition, applications are typically provisioned to handle expected peak loads. Both lead to over-provisioning and low resource utilization.John Foley wrote an...

A joint Webinar with CohesiveFT - Making Cloud Portability a practical reality

Dekel Tankel — Mon, 10 Nov 2008 18:10:00 +0000

You are welcome to join us next week – Tuesday, November 18th, 1PM EST, to a joint Webinar with our good partners – CohesiveFT. During this 1 hour Webinar we will provide a live demo illustrating how you can deploy, scale and control a complex ...

The impact of cloud computing on build vs. buy behaviour

Nati Shalom — Mon, 10 Nov 2008 15:54:18 +0000

Last week I took part in an interesting discussion with a group of architects, and the question of build vs. buy came up. It came up specifically in the context of the recent experience with many of new Internet companies....

GigaSpaces at Supercomputing 2008, Nov 17-20, Austin, TX

Amnon Raviv — Mon, 10 Nov 2008 09:45:54 +0000

Want better ROI on your HPC investment? Want to move transactional, business-critical applications to your grid and rip the benefits of the “new HPC”? Come visit GigaSpaces at the Sun booth in SC08, the international conference for high performance computing. Learn how our joint offerings help enterprises of all sizes run more complex simulations, forecasts and models [...]

Scalable, Low Latency Web Tier on Amazon EC2

Geva Perry — Fri, 07 Nov 2008 20:24:17 +0000

Shay Hassidim, deputy CTO at GigaSpaces, posted an impressive write-up of a benchmark the team ran on Amazon EC2. What's nice about it is that they took a standard web app, in this case the Spring PetClinic, and dropped it...

Scaling the Web Layer – The Web Container Benchmark

Shay Hassidim — Fri, 07 Nov 2008 05:07:54 +0000

We have been working for the past few days on building a benchmark for our new web containers and its built-in integration with the apache load balancer. The benchmark was running on Amazon EC2/EBS cloud and was deployed using our new web based cloud tools. The benchmark goal is to measure the scalability of the [...]

HelloWorld — a bonus readme

Owen Taylor — Wed, 05 Nov 2008 00:29:30 +0000

When you download the GigaSpacesXAP product (currently at version 6.6.1) you get a few really nice examples.

With those examples come some very detailed instructions and explanations as to how to use GigaSpaces. These instructions are found in the "docs" subfolder under the example root and are written in rich HTML with beautiful screenshots and graphics to help guide the reader.

Example:

GS_HOME/examples/helloworld/docs

Despite the existence of these terrific resources, for some people who are new to the GigaSpaces experience, it might be helpful to have a set of instructions in the traditional readme.txt file. ( I can be so old fashioned sometimes...)

I created the following in the hopes that it will prove useful and speed up the adoption of this wonderful application platform known as GigaSpacesXAP.

Here goes... (Imagine this readme is located in the root of the helloworld example like this:

GS_HOME/examples/helloworld/readme )

The Owen Taylor supplementary HelloWorld README for the GigaSpacesXAP6.6.1 helloworld example:

Hello and welcome to Space-based Computing!

This example shows the basic interaction between a processing unit (where objects are processed) and a simple client that feeds the processing unit with objects to be processed.

There are two ways to run this example: (provided as scripts in this directory)

The first way is simplest and shows the processing unit running in one process and the Feed running in another. Neither of the processes are managed and no clustering technology is employed to provide fail-over or scaling.

To try out the example in this simple way read and follow the first set of instructions below.

[To try out the more complex topology, read and follow the second set of instructions below.]

_______________________________________
FIRST SET OF INSTRUCTIONS::
_______________________________________

This example does the following:
1) starts the processing unit containing the processor and a space
[this is where the work of processing information goes on]
2) Starts the client-side Feeder
[this populates the system with 1000 objects to be processed]

Instructions:
1) Ensure you have GigaSpacesXAP6.6 or higher
2) Ensure you have JDK1.5
3) Navigate to the directory containing this example readme file

Execute the following
> build.sh dist
It will:

a) Build the application and create the jar file versions of the deployment units

Next Execute:

> ../../bin/puInstance.sh ./processor/pu/hello-processor.jar

This will:

b) Start the space-side system which includes a processor and a space

Edit the build.xml file so that this line:

<target name="run-feeder" depends="dist">

Is changed to equal this:

<target name="run-feeder" depends="">

Next Execute:

> build.sh run-feeder

This will:

c) Start the client-side Feeder which populates the system with 1000 objects

If you want to use a GUI to examine what is running, execute:
>../../bin/gs-ui.sh

If you want to feed more information into the system,
you can execute the

> build.sh run-feeder

more than once.

Note: If you wish to open the three Eclipse projects that make up the source for
this example in Eclipse, be aware that they use a variable GS_HOME that must be
configured in your Eclipse workspace. It should point to the install
directory/folder of GigaSpacesXAP6.5

_______________________________________
SECOND SET OF INSTRUCTIONS::
_______________________________________

This example does the following:

1) Starts the GigaSpacesXAP runtime environment (Service Grid)
2) Deploys the space-side system as a "cluster" which is split into 4 processing units:

2 partitions
(which divide the total work up between them)

2 dedicated backups
(one for each active partition)

3) Starts the client-side Feeder
[this populates the entire system with 1000 objects to be processed]

Instructions:
1) Ensure you have GigaSpacesXAP6.6 or higher
2) Ensure you have JDK1.5
3) Navigate to the directory containing this example readme file

Execute the following:
> build.sh dist
It will:

a) Build the application and create the jar file versions of the deployment units

Next Execute: (add piping to a log file if you like)

> ../../bin/gsc.sh &
> ../../bin/gsc.sh &
> ../../bin/gsm.sh &

This will:

b) Start the service grid which is the Gigaspaces application server runtime environment

Next Execute:

> ../../bin/gs-ui.sh &

This will:
c) Start the gs-ui so you can see when the service grid is started

(Switch to the middle tab on the left-hand side of the GUI called “Deployments,Details” and you should see two boxes in the bottom left of the screen. Those are your empty GSC runtime containers.)

Edit the build.xml file so that this section:

<macrodef name="deploy">
<attribute name="name"/>
<sequential>
<java classname="org.openspaces.pu.container.servicegrid.deploy.Deploy" fork="false">
<classpath refid="all-libs"/>
<arg value="-groups"/>
<arg value="${groups}"/>
<arg value="-timeout"/>
<arg value="15000"/>
<arg value="@{name}"/>
</java>
</sequential>
</macrodef>

Is changed to equal this:

<macrodef name="deploy">
<attribute name="name"/>
<sequential>
<java classname="org.openspaces.pu.container.servicegrid.deploy.Deploy" fork="false">
<classpath refid="all-libs"/>
<arg value="-groups"/>
<arg value="${groups}"/>
<arg value="-timeout"/>
<arg value="90000"/>
<arg value="-locators"/>
<arg value="localhost"/> 
<arg value="-cluster"/>
<arg value="schema=partitioned-sync2backup"/>
<arg value="total_members=2,1"/>
<arg value="@{name}"/>
</java>
</sequential>
</macrodef>

This edit changes the topology of the application being deployed so that there will be 2 primary active instances and 2 backup instances deployed into the service grid.

Next Execute:

> build.sh deploy-processor

This will deploy the newly re-defined 2 primary spaces and accompanying workers and their backup service instances into the service grid (you will see them appear in the gs-ui GUI)

Wait for all the nodes to appear. (There should be 2 nodes with a ‘p’ for primary and 2 nodes with a ‘b’ for backup)
This can take a while… (up to 2 minutes the first time depending on the machine and network)

Next Execute:

> build.sh run-feeder

This will:

c) Start the client-side Feeder which populates the system with 1000 objects

If you want to feed more information into the system,
you can execute the

> build.sh run-feeder

more than once.

Note: If you wish to open the three Eclipse projects that make up the source for
this example in Eclipse, be aware that they use a variable GS_HOME that must be
configured in your Eclipse workspace. It should point to the install
directory/folder of GigaSpacesXAP6.6
HTH Owen.

Cloud Computing. Literally.

Geva Perry — Mon, 03 Nov 2008 19:19:35 +0000

Last week we made a very exciting announcement about Miwok Airways selecting GigaSpaces as the application server for running their reservation and pricing engine which will run on EC2. This is a great case study for cloud computing. For one...

The GigaSpaces Blog