<html><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; "><div><font class="Apple-style-span" face="Times">Tim:</font></div><div><font class="Apple-style-span" face="Times"><br></font></div><div><font class="Apple-style-span" face="Times">Yes, running experiments is going to be an important early task.</font></div><div><font class="Apple-style-span" face="Times"><br></font></div><div><font class="Apple-style-span" face="Times">Ian.</font></div><div><br></div><br><div><div>On May 12, 2009, at 11:53 AM, Tim Freeman wrote:</div><br class="Apple-interchange-newline"><blockquote type="cite"><div>On Tue, 12 May 2009 11:42:00 -0500<br>yizhu <<a href="mailto:yizhu@cs.uchicago.edu">yizhu@cs.uchicago.edu</a>> wrote:<br><br><blockquote type="cite">That's great. It makes file management between EC2 and S3 much easier, I <br></blockquote><blockquote type="cite">will definitely check it out.<br></blockquote><blockquote type="cite"><br></blockquote><br>Note that every I/O op to S3 will be significantly slower than EBS (and<br>probably also slower than the local scratch disk). Add in FUSE (userspace) and<br>that only compounds the problem. So if you use this option I would suggest<br>taking measurements to make sure it is acceptable for the task at hand.<br><br>Tim<br><br><blockquote type="cite">Michael Wilde wrote:<br></blockquote><blockquote type="cite"><blockquote type="cite">Maybe relevant - there's a FUSE filesystem to mount an S3 bucket as a <br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">filesystem:<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><a href="http://code.google.com/p/s3fs/wiki/FuseOverAmazon">http://code.google.com/p/s3fs/wiki/FuseOverAmazon</a><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">- Mike<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">On 5/12/09 11:19 AM, Ioan Raicu wrote:<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">Hi,<br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">yizhu wrote:<br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">Hi,<br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">I've now got the swift system running on Amazon EC2 with swift <br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">installed on head-node of the cluster. I will soon let it the user <br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">submit the job from its own machine after i solve the authentication <br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">issues of Globus.<br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">I think my next step is to write a sample swift code to check if we <br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">can let swift grab input files from S3, execute it, and then write <br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">the output files back to S3.<br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">Since each Amazon virtual node only has limited storage space( Small <br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">instance for 1.7GB, Large Instance for 7.5GB), <br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">Are you sure? That sounds like like the amount of RAM each instance <br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">gets. The last time I used EC2 (more than a year ago), each instance <br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">had a disk of 100GB+ each, which could be treated as a scratch disk <br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">that would be lost when the instance was powered down.<br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">we may need to use EBS(Elastic Block Store) to storage temp files <br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">created by swift. The EBS behaved like a hard disk volume and can be <br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">mounted by any virtual nodes. But here arise a problem, since a <br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">volume can only be attached to one instance at a time[1], so the <br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">files store in EBS mounted by one node won't be shared by any other <br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">nodes; <br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">The EBS sounds like the same thing as the local disk, except that it <br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">is persisted in S3, and can be recovered when the instance is started <br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">again later.<br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">You can use these local disks, or EBS disks, to create a <br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">shared/parallel file system, or manage them yourself.<br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">we lost the file sharing ability which I think is a fundamental <br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">requirement for swift.<br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">The last time I worked with the Workspace Service (now part of <br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">Nimbus), we were able to use NFS to create a shared file system across <br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">our virtual cluster. This allowed us to run Swift without <br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">modifications on our virtual cluster. Some of the more recent work on <br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">Swift might have eased up some of the requirements for a shared file <br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">system, so you might be able to run Swift without a shared file <br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">system, if you configure Swift just right. Mike, is this true, or we <br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">aren't quite there yet?<br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">Ioan<br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">-Yi<br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">Michael Wilde wrote:<br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">On 5/7/09 12:54 PM, Tim Freeman wrote:<br></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">On Thu, 07 May 2009 11:39:40 -0500<br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">Yi Zhu <<a href="mailto:yizhu@cs.uchicago.edu">yizhu@cs.uchicago.edu</a>> wrote:<br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">Michael Wilde wrote:<br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">Very good!<br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">Now, what kind of tests can you do next?<br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">Next, I will try to let swift running on Amazon EC2.<br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">Can you exercise the cluster with an interesting workflow?<br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">Yes, Is there any complex sample/tools i can use (rahter than <br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">first.swift) to test swift performance? Is there any benchmark <br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">available i can compare with?<br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">How large of a cluster can you assemble in a Nimbus workspace ?<br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">Since the vm-image i use to test 'swift' is based on NFS shared <br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">file system, the performance may not be satisfiable if the we have <br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">a large scale of cluster. After I got the swift running on Amazon <br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">EC2, I will try to make a dedicate vm-image by using GPFS or any <br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">other shared file system you recommended.<br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">Can you aggregate VM's from a few different physical clusters <br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">into one Nimbus workspace?<br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">I don't think so. Tim may make commit on it.<br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">There is some work going on right now making auto-configuration <br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">easier to do<br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">over multiple clusters (that is possible now, it's just very <br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">'manual' and<br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">non-ideal unlike with one physical cluster). You wouldn't really <br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">want to do NFS<br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">across a WAN, though.<br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">Indeed. Now that I think this through more clearly, one workspace == <br></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">one cluster == one Swift "site", so we could aggregate the resources <br></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">of multiple workspaces through Swift to execute a multi-site workflow.<br></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">- Mike<br></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">What's the largest cluster you can assemble with Nimbus?<br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">I am not quite sure,I will do some test onto it soon. since it is <br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">a EC2-like cloud, it should easily be configured as a cluster with <br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">hundreds of nodes. Tim may make commit on it.<br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">I've heard of EC2 deployments in the 1000s at once, it's up to your <br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">EC2 account<br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">limitations (they seem pretty efficient with making your quota <br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">ever-higher).<br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">Nimbus installation at Teraport maxes out at 16, there are other <br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">'science<br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">clouds' but I don't know their node numbers. EC2 is the place <br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">where you will<br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">be able to really test scaling something.<br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite">Tim<br></blockquote></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote></blockquote><blockquote type="cite"><blockquote type="cite"><br></blockquote></blockquote><blockquote type="cite"><br></blockquote></div></blockquote></div><br></body></html>