Data analysis on cloud

[rob9999]

 Hi,

I have some datafiles that I have to store in a cloud drive run by the university (it has a https:// address). I have mapped that drive onto my device under "A:\". 

I have managed to run SAS analysis on these files without downloading them to the local device. It does run a lot slower, but I am fortunate that the sample size are not huge. I am not able to download the file onto my local device for security reason. 

May I know, when I run the analysis, will SAS9.4 create a temp file on local device, e.g. analysis output? If so, where can I locate them? Can I limit all relating files to be saved on the cloud drive only? 

I understand that my question is very confusing as I am not very familiar with the these computer terminology. Ultimately, the goal is not to leave any trails of the data on my local device. I am running these on Windows 7

Thank you.

[Reeza]

What type of SAS installation are you using? SAS Academics on Demand is cloud based, as is SAS UE on AWS. 

SAS UE that you download runs locally on your machine. The answer to your question depends heavily on your SAS installation set up. 

In general though:

May I know, when I run the analysis, will SAS9.4 create a temp file on local device, e.g. analysis output? If so, where can I locate them? Can I limit all relating files to be saved on the cloud drive only? 

SAS will likely create temp files on your local drive in either your WORK library or a user library. You can set this to be located on the A drive as well. Output can go on either the local or user drive, but it's generally considered aggregate at that point and you have different restrictions. Again, how you access these files depends on your installation.

---

@rob9999 wrote:

 Hi,

I have some datafiles that I have to store in a cloud drive run by the university (it has a https:// address). I have mapped that drive onto my device under "A:\". 

I have managed to run SAS analysis on these files without downloading them to the local device. It does run a lot slower, but I am fortunate that the sample size are not huge. I am not able to download the file onto my local device for security reason. 

May I know, when I run the analysis, will SAS9.4 create a temp file on local device, e.g. analysis output? If so, where can I locate them? Can I limit all relating files to be saved on the cloud drive only? 

I understand that my question is very confusing as I am not very familiar with the these computer terminology. Ultimately, the goal is not to leave any trails of the data on my local device. I am running these on Windows 7

Thank you.

---

[rob9999]

Hi,

Reeza thank you for your comment. My SAS setup is SAS base 9.4 for Windows, not the cloud version. May I know how do I setup so that all output and work files will be saved on the A:?
Thank you.

[Reeza]

If you're on SAS 9.4 and not University Edition, you have a paid license through either academia or a company. If it's a company you can call tech support and get a verified answer. I'll move this to admin/set up so the appropriate users on here can answer. 

If you want to set it up permanently that's different than setting it up for a specific project. 

If it's a specific project, I would set the USER library which directs all files to the location specified - this will slow down your analysis. I also don't know where the temporary results files would be saved, which is why I'm moving it the Admin sub-forum.

If it's permanent, I suspect you need to make changes to your config file to set it up so that the work library and other files are sent to the desired location. 

libname user 'A:\SAS\Temp\';

---

@rob9999 wrote:
Hi,

Reeza thank you for your comment. My SAS setup is SAS base 9.4 for Windows, not the cloud version. May I know how do I setup so that all output and work files will be saved on the A:?
Thank you.

---

[rob9999]

Hi Reeza,

Thank you for your advise.

[ybolduc · Accepted solution]

Hi Rob,

The way you describe your situation triggers a lot of alarms in my security conscious mind.

You may also want to consider the rational of not storing the data on your computer. If you can't do this because the data is not allowed to cross a territorial border (like should stay in the US), then by having the processing happen on your machine you may be violating that restriction. Even if you don't store the data on a persistent storage, you still have to take it to your computer's RAM.

You should also consider the fact that Windows actually does swapping in some cases, so even if you move you assign your work to the A:\ drive, you may still have unknowingly written the data to your local drive.

I'm just bringing these up because if you actually signed an agreement to not take the data out of the cloud or out of a territory, you may actually be violating that agreement by having the processing happen on your machine.

Thanks,

[rob9999]

Thanks ybolduc. 

Yes, the data I am dealing require secure access and storage, as it is bound by the privacy act. The main reason that we don't keep data on local is to avoid risk of pc/laptop being stolen. However, there is no restriction that the data cannot cross territories etc.

It just that we need to keep all files in one location to so that all users will be working on same version and avoiding data to spread all over places. Also, there's an expiry date on usage period, i.e. we have to delete all data after 10 years (including any possible backup version--that's another potential issue with cloud), so it is easier to delete the single folder where all data are stored. 

    "Even if you don't store the data on a persistent storage, you still have to take it to your computer's RAM."

Yes, your statement confirms my suspicion that temp file will be created while I run the analysis on the files stored remotely. 

I might have to look into alternative solution.