[IPython-dev] ipython html notebook server on dotcloud

Brian Granger ellisonbg at gmail.com
Wed Sep 5 18:58:22 EDT 2012


Hi, these are things we are interested in.  There are a number of
different issues you are bringing up:

* Making the notebook server multiuser aware.  This is definitely on
our radar and we are trying to get this work funded through academic
grants.  We are at the early stages of the planning of this.
* Running the notebook server on PaaS platforms.  This is an entirely
separate issue that we will have to revisit in the future - stay tuned
for more on this in the coming months.  One issue you will run into
though is that we rely on zeromq/pyzmq which definitely doesn't run on
most of the Paas vendors I know of.



On Wed, Sep 5, 2012 at 6:09 AM, Shoibal Chakravarty <shoibalc at gmail.com> wrote:
> Hi,
> I just got interested in the possibility of getting ipython html notebook
> server on the various Platform as a Service (PaaS) ideas out there:
> Openshift, dotcloud, heroku, appfog etc. I was interested in seeing if I
> could get people to try out my other python projects  [energy economics
> modeling] without having to take the trouble of downloading and compiling a
> lot of different packages. I would put everything required on one of these
> free PaaS services and have them try it out via the ipython notebook
> interface.
> Carl Smith's  notebookcloud (http://github.com/carlsmith) was another
> inspiration. The PaaS approach is a baby version of  the AMI approach of
> notebookcloud. I let the PaaS vendors take care of the building and
> maintaining the AWS infrastructure. It also happens to be completely free
> for small applications.
> Most of the Paas services do not support websockets and and often expect
> only wsgi compliant python web servers. Only  dotcloud
> (http://www.dotcloud.com/) seems to work for now with tornado and
> websockets. See http://github.com/shoibalc/ipynbserver-dotcloud.
> I am still concerned about security implications of the notebook server.
> Ideally, I would like  the following:
> 1. Let anybody login with a login/passwd or the various authentication
> services.
> 2. Give them a secure walled-in temporary directory to work in, with
> restricted shell function access.
> 3. Limit resources they use (memory, CPU load etc) and automatically log
> them out if the violate this.
> 4. Limit sessions to xx minutes.
> 5. Set up a parallel database service for authentication and to let users
> save their notebooks etc. (eg, MongoDB uses json as storage)
> Which of these would be easily feasible? Especially point 2. Sage seems to
> create (and later destroy) a temporary working directory for each cell  and
> severely limits shell access. One directory per login with limited outside
> access might be better.
> Please let me know what ideas/suggestions you have. Thanks a lot for
> ipython!
> -Shoibal.
> _______________________________________________
> IPython-dev mailing list
> IPython-dev at scipy.org
> http://mail.scipy.org/mailman/listinfo/ipython-dev

Brian E. Granger
Cal Poly State University, San Luis Obispo
bgranger at calpoly.edu and ellisonbg at gmail.com

More information about the IPython-dev mailing list