Aggregate on AWS

Thanks to Chris Robert, we now have some documentation on setting up
ODK Aggregate to run on a Linux micro-instance on the Amazon Web
Services EC2 infrastructure. You can find the details at
http://code.google.com/p/opendatakit/wiki/AggregateAWSInstall

As always, if you have any documentation that would be useful to the
community, we are glad to link to (or post) it.

Good stuff....

··· On Mon, May 14, 2012 at 1:45 PM, Yaw Anokwa wrote:

Thanks to Chris Robert, we now have some documentation on setting up
ODK Aggregate to run on a Linux micro-instance on the Amazon Web
Services EC2 infrastructure. You can find the details at
http://code.google.com/p/opendatakit/wiki/AggregateAWSInstall

As always, if you have any documentation that would be useful to the
community, we are glad to link to (or post) it.

--
Post: opendatakit@googlegroups.com
Unsubscribe: opendatakit+unsubscribe@googlegroups.com
Options: http://groups.google.com/group/opendatakit?hl=en

--
Play it forward and Have a good one!

"Like a camera I use the negative to develop..uDig"
"Success is the ability to go from one failure to another with no loss of
enthusiasm"
---------------------------------------------------------------------------------------------------------->
OutSourceRox*
*
Roxroy K. Bollers
GIS and IT Consultant
tel. 592.685.2737
.........................................................
Follow Us On
Facebook http://www.facebook.com/OutSourceRox
Google+ https://plus.google.com/114064153117817313312
Web http://twitinternational.com/outsourcerox.htm

How does the availability and response time compare to AppEngine?? Curious!!

··· On Monday, May 14, 2012 8:45:12 PM UTC+3, Yaw Anokwa wrote: > Thanks to Chris Robert, we now have some documentation on setting up > ODK Aggregate to run on a Linux micro-instance on the Amazon Web > Services EC2 infrastructure. You can find the details at > http://code.google.com/p/opendatakit/wiki/AggregateAWSInstall > > As always, if you have any documentation that would be useful to the > community, we are glad to link to (or post) it.

Unlike GAE, in the AWS framework your instance is basically always running
-- so I don't think that there is the same fire-up lag that you sometimes
get with GAE. Overall compute power and performance during heavy usage will
depend on the instance type (see
http://docs.amazonwebservices.com/AWSEC2/latest/UserGuide/concepts_micro_instances.htmlfor
info on the free micro instances and
http://aws.amazon.com/ec2/instance-types/ for info on other instance
types). The micro instance has performed well for me during testing, but I
don't know whether I'd bump into limits during a large-scale deployment; I
suspect that a micro instance may be enough for even moderately heavy
usage, but I can't be sure. Certainly a "standard instance" should perform
perfectly well, for basically a fixed monthly cost.

One option is to pilot with a micro instance and then upgrade to a standard
instance just during the time period of heavy deployment, then downgrade
back to a micro instance. You could even try deploying on the micro
instance and stand ready to upgrade mid-deployment if there is any
performance issue.

It will take me a few months to deploy anything, and even then I may be
prudent and just go with a standard instance since I'll be deploying
multiple clients/projects on a single instance. If anybody else learns
anything about AWS performance in the mean-time, please do share!

Thanks,

Chris

··· On Friday, May 18, 2012, wrote:

On Monday, May 14, 2012 8:45:12 PM UTC+3, Yaw Anokwa wrote:

Thanks to Chris Robert, we now have some documentation on setting up
ODK Aggregate to run on a Linux micro-instance on the Amazon Web
Services EC2 infrastructure. You can find the details at
http://code.google.com/p/opendatakit/wiki/AggregateAWSInstall

As always, if you have any documentation that would be useful to the
community, we are glad to link to (or post) it.

How does the availability and response time compare to AppEngine??
Curious!!

--
Post: opendatakit@googlegroups.com <javascript:;>
Unsubscribe: opendatakit+unsubscribe@googlegroups.com <javascript:;>
Options: http://groups.google.com/group/opendatakit?hl=en

Since most of the time doing surveys is spent gathering data or in transit
to the next interviewee, and since each phone will send all of its
submissions one-at-a-time to the server, unless you are running very large
surveys with 100's of ODK Collect devices clamoring to upload data all at
the same time (e.g., at the very end of the day), it is unlikely the
servers will ever run much above idle. If you have no media attachments in
those submissions, you can likely bump that number up to the 10,000's of
devices. And if submissions steadily trickle in over the course of the
day, I cannot envision any survey short of one on a national scale
producing any significant load on the server.

With Tomcat, you should consider increasing the socket and request timeout
limits to handle slow satellite uplink or other issues your field personnel
might experience.

Of course, you will likely need to tweak the default settings for the
Tomcat, Java and MySQL systems to tune them to handle the larger workloads,
but the AWS Micro instance should be able to handle it, and certainly the
Small instance.

Mitch

··· On Thu, May 17, 2012 at 9:44 PM, Christopher Robert wrote:

Unlike GAE, in the AWS framework your instance is basically always running
-- so I don't think that there is the same fire-up lag that you sometimes
get with GAE. Overall compute power and performance during heavy usage will
depend on the instance type (see
http://docs.amazonwebservices.com/AWSEC2/latest/UserGuide/concepts_micro_instances.htmlfor info on the free micro instances and
http://aws.amazon.com/ec2/instance-types/ for info on other instance
types). The micro instance has performed well for me during testing, but I
don't know whether I'd bump into limits during a large-scale deployment; I
suspect that a micro instance may be enough for even moderately heavy
usage, but I can't be sure. Certainly a "standard instance" should perform
perfectly well, for basically a fixed monthly cost.

One option is to pilot with a micro instance and then upgrade to a
standard instance just during the time period of heavy deployment, then
downgrade back to a micro instance. You could even try deploying on the
micro instance and stand ready to upgrade mid-deployment if there is any
performance issue.

It will take me a few months to deploy anything, and even then I may be
prudent and just go with a standard instance since I'll be deploying
multiple clients/projects on a single instance. If anybody else learns
anything about AWS performance in the mean-time, please do share!

Thanks,

Chris

On Friday, May 18, 2012, wrote:

On Monday, May 14, 2012 8:45:12 PM UTC+3, Yaw Anokwa wrote:

Thanks to Chris Robert, we now have some documentation on setting up
ODK Aggregate to run on a Linux micro-instance on the Amazon Web
Services EC2 infrastructure. You can find the details at
http://code.google.com/p/opendatakit/wiki/AggregateAWSInstall

As always, if you have any documentation that would be useful to the
community, we are glad to link to (or post) it.

How does the availability and response time compare to AppEngine??
Curious!!

--
Post: opendatakit@googlegroups.com
Unsubscribe: opendatakit+unsubscribe@googlegroups.com
Options: http://groups.google.com/group/opendatakit?hl=en

--
Post: opendatakit@googlegroups.com
Unsubscribe: opendatakit+unsubscribe@googlegroups.com
Options: http://groups.google.com/group/opendatakit?hl=en

--
Mitch Sundt
Software Engineer
University of Washington
mitchellsundt@gmail.com