RapidScale Clusters, LLC

Amazon EC2 Dual Xeon Virtual Render Node Pricing Plan

This product is in Beta. It may not work perfectly at all times, but we will support you 100% during our efforts to complete beta testing and finalize production quality code. Thank you for your supporting us in our efforts to bring you the very best and most cutting edge render farm technology on the market!

Product Description: 

An Amazon Machine Image, or "virtual server", that joins into your local render farm running RenderFarmer (from www.linuxrenderfarm.com) and immediately adds to your compute capacity. Charges are per hour. All traffic flows through an encrypted VPN. The number of virtual nodes--or instances--that can be launched is theoretically unlimited, but in reality depends on bandwidth, the speed of your server running RenderFarmer, and other practical considerations.

 

Terms and Conditions: 

This item is provided by RapidScale Clusters LLC. There is a $25 setup fee that entitles you to a customized AMI that is used for virtual render nodes. This AMI is tailored with specific settings that allow it to connect to your network and participate in your render farm, just as if it were on your local network. This product only works in conjunction with RenderFarmer from RapidScale Clusters. An Amazon EC2 account is required.  

Use of a render node incurs a per hour, per instance charge that includes Amazon's usage fees. There is no contract, no monthly charge, and no obligation. **Charges accrue until the virtual render nodes are powered off!** You MUST power off the render nodes when you no longer wish to actively use them. See example below.

Amazon Web Services will monitor usage and bill you directly for the time and bandwidth consumed by the render node instances you start and run. RapidScale Clusters cannot reimburse you for Amazon's charges accrued by nodes that are not working correctly. You MUST ensure that your renders are working as planned and terminate the virtual machine instances if they are not. It is possible that Internet outages, problems on your local network, etc. could cause the render nodes to be "orphaned" and remain running but not processing any jobs. For instance, if you accidentally stop the local VPN service, the EC2 nodes will not be able to see render jobs. We make every effort to automatically detect these situations in software and power off the nodes automatically, but ultimately the responsibility for accrued charges is yours. Fortunately, it is easy to see connected nodes in the render queue manager and determining if the nodes are working is generally very obvious. 

 

Usage Charges - United States

Usage Charges - European Union

Example:  

I have a 1000 frame animation that I would like to render. I submit the job inside RenderFarmer, which is idle because I have no local nodes in my render farm. If I did have some local nodes, they would begin working on the job immediately. 

Next I log in to my Amazon EC2 Console at https://console.aws.amazon.com/ec2/home and launch 20 instances of my Dual Xeon Render Node from RapidScale Clusters. Within about 5 minutes the nodes have finished booting, established their VPN tunnels to my network, and joined into the render farm as compute nodes. Now they will take jobs from the render queue. Under the "Computers" tab in the DrQueue Render Queue Manager, I can see that 20 EC2 render nodes--40 2.0 Ghz processors total--are connected and working in addition to any local nodes I may have.

One hour and 50 minutes later, the render job is done and I would like to shut off the render nodes and stop the per hour charges. Once again I log into my Amazon EC2 Console at https://console.aws.amazon.com/ec2/home and click on the "Instances" link. The twenty Dual Xeon Render Node instances are listed there with a status of "Running". I select the check box next to each node and then click the "Terminate" button. The status of each node immediately changes to "Shutting down" and within a minute or two reads "Terminated". 

My charges for this job are calculated as follows. Twenty (20) nodes were used for 1.83 hours. Amazon rounds up partial hours so this will be treated as 20 nodes x 45 cents per hour x 2 hours, for a subtotal of $18.00 USD. Also, I transferred the 1000 images rendered so Amazon will bill a small amount for data transfer at 17 cents per gigabyte. This may be an additional 20 cents. At the end of this month, Amazon Web Services will bill me for $18.20 USD.

In summary, I paid about nine dollars per hour for the twenty dual xeon render nodes I "rented". They joined into my render farm immediately and went to work on my project, which completed in less than 2 hours instead of 24 - 40 hours. If I had a few local nodes as well, it would have completed even faster and my Amazon EC2 charges may have been smaller. The EC2 Dual Xeon Render Node gave me the ability to grow my render farm on demand by any number of nodes and then shrink it back down again when I was done, without buying, configuring, or paying the electric / cooling bill for a single piece of hardware.

Sign Up Now!