You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@cmda.apache.org by Qihao Bao <qi...@cmu.edu> on 2016/07/29 22:28:51 UTC
User Mapping between JPL_backend and CMU_Provenance
Hi all,
*I have thought about the user mapping issue based on Benyang's suggestion
yesterday*. As we know, each user would be assigned to a VM instance.
*Lei,*
I think you can get the instance URL inside of the entry.py? For example,
you can get http://ec2-52-53-198-227.us-west-1.compute.amazonaws.com/
dynamically
for one instance and another URL for another instance, right?
So if you add a dict of 'I*nstance URL to UserId*' and still send UserId to
CMU_provenance as identification, everything would still work and nobody
else would have to change their code.
What i can do is to provide you the mapping of 'UserName to UserID in
CMU_provenance', since you already have the mapping of ' instance URL to
UserName', you will be able to join them together.
Thanks a lot.
Qihao
Re: User Mapping between JPL_backend and CMU_Provenance
Posted by "Pan, Lei (398K)" <le...@jpl.nasa.gov>.
Hi Qihao:
Your point is taken.
I will start adding such a dict in entry.py.
Thanks,
-Lei
On 8/2/16, 4:14 PM, "Qihao Bao" <qi...@cmu.edu> wrote:
>Lei,
>
>I think CMU_frontend is able to get user info through cookie.
>The issue is that entry.py needs to know who is the user. To realize that,
>the user info should have to be passed through a long chain:
>
>CMU_frontend --->> Benyang's web page -->> entry.py -->> and then finally
>CMU_backend.
>
>Changing so much code might be risky at this stage. That's why i think the
>method i mentioned before should be easier.
>
>Thanks
>
>Br.
>Qihao
>
>
>On Sat, Jul 30, 2016 at 2:03 AM, Pan, Lei (398K) <le...@jpl.nasa.gov>
>wrote:
>
>> Hi Qihao:
>>
>> The idea of identifying a user from a pre-defined, hard-coded
>> url to username map would work. However a more automatic way
>> is for a client to find out who he/she is after logging on to
>> the provenance server. Is there a cookie or query mechanism
>> that we can leverage? The login mechanism you use may provide
>> some programming interface.
>>
>> Thanks,
>> -Lei
>>
>> On 7/29/16, 3:28 PM, "Qihao Bao" <qi...@cmu.edu> wrote:
>>
>> >Hi all,
>> >
>> >*I have thought about the user mapping issue based on Benyang's
>>suggestion
>> >yesterday*. As we know, each user would be assigned to a VM instance.
>> >
>> >*Lei,*
>> >
>> >I think you can get the instance URL inside of the entry.py? For
>>example,
>> >you can get http://ec2-52-53-198-227.us-west-1.compute.amazonaws.com/
>> >dynamically
>> >for one instance and another URL for another instance, right?
>> >
>> >So if you add a dict of 'I*nstance URL to UserId*' and still send
>>UserId
>> >to
>> >CMU_provenance as identification, everything would still work and
>>nobody
>> >else would have to change their code.
>> >
>> >What i can do is to provide you the mapping of 'UserName to UserID in
>> >CMU_provenance', since you already have the mapping of ' instance URL
>>to
>> >UserName', you will be able to join them together.
>> >
>> >Thanks a lot.
>> >
>> >
>> >Qihao
>>
>>
Re: User Mapping between JPL_backend and CMU_Provenance
Posted by Jia Zhang <ji...@sv.cmu.edu>.
I tend to agree with Qihao at this point, from a conservative perspective
for the summer school.
How would you think Lei?
Best regards,
Jia
On Tue, Aug 2, 2016 at 4:14 PM, Qihao Bao <qi...@cmu.edu> wrote:
> Lei,
>
> I think CMU_frontend is able to get user info through cookie.
> The issue is that entry.py needs to know who is the user. To realize that,
> the user info should have to be passed through a long chain:
>
> CMU_frontend --->> Benyang's web page -->> entry.py -->> and then finally
> CMU_backend.
>
> Changing so much code might be risky at this stage. That's why i think the
> method i mentioned before should be easier.
>
> Thanks
>
> Br.
> Qihao
>
>
> On Sat, Jul 30, 2016 at 2:03 AM, Pan, Lei (398K) <le...@jpl.nasa.gov>
> wrote:
>
> > Hi Qihao:
> >
> > The idea of identifying a user from a pre-defined, hard-coded
> > url to username map would work. However a more automatic way
> > is for a client to find out who he/she is after logging on to
> > the provenance server. Is there a cookie or query mechanism
> > that we can leverage? The login mechanism you use may provide
> > some programming interface.
> >
> > Thanks,
> > -Lei
> >
> > On 7/29/16, 3:28 PM, "Qihao Bao" <qi...@cmu.edu> wrote:
> >
> > >Hi all,
> > >
> > >*I have thought about the user mapping issue based on Benyang's
> suggestion
> > >yesterday*. As we know, each user would be assigned to a VM instance.
> > >
> > >*Lei,*
> > >
> > >I think you can get the instance URL inside of the entry.py? For
> example,
> > >you can get http://ec2-52-53-198-227.us-west-1.compute.amazonaws.com/
> > >dynamically
> > >for one instance and another URL for another instance, right?
> > >
> > >So if you add a dict of 'I*nstance URL to UserId*' and still send UserId
> > >to
> > >CMU_provenance as identification, everything would still work and nobody
> > >else would have to change their code.
> > >
> > >What i can do is to provide you the mapping of 'UserName to UserID in
> > >CMU_provenance', since you already have the mapping of ' instance URL to
> > >UserName', you will be able to join them together.
> > >
> > >Thanks a lot.
> > >
> > >
> > >Qihao
> >
> >
>
Re: User Mapping between JPL_backend and CMU_Provenance
Posted by Qihao Bao <qi...@cmu.edu>.
Lei,
I think CMU_frontend is able to get user info through cookie.
The issue is that entry.py needs to know who is the user. To realize that,
the user info should have to be passed through a long chain:
CMU_frontend --->> Benyang's web page -->> entry.py -->> and then finally
CMU_backend.
Changing so much code might be risky at this stage. That's why i think the
method i mentioned before should be easier.
Thanks
Br.
Qihao
On Sat, Jul 30, 2016 at 2:03 AM, Pan, Lei (398K) <le...@jpl.nasa.gov>
wrote:
> Hi Qihao:
>
> The idea of identifying a user from a pre-defined, hard-coded
> url to username map would work. However a more automatic way
> is for a client to find out who he/she is after logging on to
> the provenance server. Is there a cookie or query mechanism
> that we can leverage? The login mechanism you use may provide
> some programming interface.
>
> Thanks,
> -Lei
>
> On 7/29/16, 3:28 PM, "Qihao Bao" <qi...@cmu.edu> wrote:
>
> >Hi all,
> >
> >*I have thought about the user mapping issue based on Benyang's suggestion
> >yesterday*. As we know, each user would be assigned to a VM instance.
> >
> >*Lei,*
> >
> >I think you can get the instance URL inside of the entry.py? For example,
> >you can get http://ec2-52-53-198-227.us-west-1.compute.amazonaws.com/
> >dynamically
> >for one instance and another URL for another instance, right?
> >
> >So if you add a dict of 'I*nstance URL to UserId*' and still send UserId
> >to
> >CMU_provenance as identification, everything would still work and nobody
> >else would have to change their code.
> >
> >What i can do is to provide you the mapping of 'UserName to UserID in
> >CMU_provenance', since you already have the mapping of ' instance URL to
> >UserName', you will be able to join them together.
> >
> >Thanks a lot.
> >
> >
> >Qihao
>
>
Re: User Mapping between JPL_backend and CMU_Provenance
Posted by "Pan, Lei (398K)" <le...@jpl.nasa.gov>.
Hi Qihao:
The idea of identifying a user from a pre-defined, hard-coded
url to username map would work. However a more automatic way
is for a client to find out who he/she is after logging on to
the provenance server. Is there a cookie or query mechanism
that we can leverage? The login mechanism you use may provide
some programming interface.
Thanks,
-Lei
On 7/29/16, 3:28 PM, "Qihao Bao" <qi...@cmu.edu> wrote:
>Hi all,
>
>*I have thought about the user mapping issue based on Benyang's suggestion
>yesterday*. As we know, each user would be assigned to a VM instance.
>
>*Lei,*
>
>I think you can get the instance URL inside of the entry.py? For example,
>you can get http://ec2-52-53-198-227.us-west-1.compute.amazonaws.com/
>dynamically
>for one instance and another URL for another instance, right?
>
>So if you add a dict of 'I*nstance URL to UserId*' and still send UserId
>to
>CMU_provenance as identification, everything would still work and nobody
>else would have to change their code.
>
>What i can do is to provide you the mapping of 'UserName to UserID in
>CMU_provenance', since you already have the mapping of ' instance URL to
>UserName', you will be able to join them together.
>
>Thanks a lot.
>
>
>Qihao