Cloud storage is a thing that makes life easier. If you are working on different computers (e.h. office, home, laptop, etc.), the cloud ensures that your data is getting synchronized accordingly without having the need to play around with USB storage devices, etc. In addition, it gives you some kind of instant fire-proof backup, since your data is always stored on at least two computers (the one you are working with + the cloud server). So the idea of a cloud is very nice in principle.
However, since Snowden told the world about the NSA, we are all very much more concerned about the security of our data. Big-company solutions (Dropbox, Google, Amazon, ...) are thus to be excluded from these thoughts immediately. Therefore, I tried a lot of self-hosted solutions over the last weeks. Some of them seem fairly crappy, but in the end I found one very nice software, which I intend to use for this purpose: the open-source project "Seafile"
It encrypts your data on the way from your computer to the server using AES-256/CBC and SSL. On the server they are stored in a git-like structure (at least that's what it appears to me), so no plain file storage on the HDD, which should be a good thing.
I have been working some weeks with seafile running on a Raspberry Pi on my desktop at home. The good thing is that I have full control and that I really know where my data is actually stored. However, the Raspberry (and also my Internet connection) is really slow and therefore I moved to a vServer in a big German hosting data center. Germans take data security very serious and I feel much more safe having them here, rather than somewhere in the US, etc. (have I said that I am German?). This system is already up and running.
Having introduced the problem: How secure is this configuration Seafile on a German vServer? There is the possibility to additionally encrypt the data client-side using a password. This might add rise the security level, but this might also add additional discomfort, when forced to type in the password from time to time for a lot of data sets (called "libraries" in Seafile).
Aucun commentaire:
Enregistrer un commentaire