I successfully managed to create a SQLite3 DB with Perl using Perl::DBI module.
I was wondering if there was a way to add encryption to the database to my existing Perl code ?
I read thoroughly the 2 following links :
Password Protect a SQLite DB. Is it possible?
SQLite with encryption/password protection
but the provided examples seem only to include proprietary software or C# code (especially this bit here https://stackoverflow.com/a/24349415/3186538).
Thanks in advance.
Well, you could run your data through any of the Crypt::* modules (::DES, ::Blowfish, ::IDEA, etc, in conjunction with ::CBC), then possibly encode it with base64 to get text, before writing it to the DB. And, of course, reverse the operation when reading. You could even create a Perl::DBICrypt module that sat above Perl::DBI and did this automagically.
However, it depends pretty much on how you're going to use it. If you're just worried about someone stealing and using your data, the encryption would be feasible since, without the key, it would be useless.
On the other hand, if you're trying to protect data in a system you distribute, then the key will be available to the attacker (since, without it, your code won't work). So encrypting in that case would be a minor inconvenience at best.
It's something that could only really work if you kept the key away from the attacker (such as if the Perl code runs in an app server controlled by you).
Basically any solution that decrypts data on a box accessible to an attacker will be vulnerable.
Related
Is there a decent database for nodejs that is able to be embeded & encrypted so that if someone does get that file of data, a password also has to be known ?
I am coming up short & though things like nosql, nedb, etc. exists, none of them see to be able to handle encryption of said data.
sqlcipher can be built for mapbox/sqlite3
Also, using encryption with nedb is really easy, you just create a crypto cipher and perform the operation in and out. See this example from bitcrypt
You'd basically do the same thing as cipherHelper and decipherHelper, but you'd put them in the afterSerialization and beforeDeserialization callbacks.
I'm building a Windows Store application that uses SQLite for data storage. I have found out, that the database is easily accessible through User's local folder (actually all apps have all data publicly exposed). Is there a way to at least weakly protect the database from access?
you need to look at ProtectedData class
http://msdn.microsoft.com/en-us/library/windows/apps/windows.security.cryptography.dataprotection.dataprotectionprovider.aspx
It exposes easy to use Protect / Unprotect methods that can be used to encrypt / decrypt that at app level. Encrypt data before writing to db and unencrypt before consuming
I also looking for the same solution and found sqlite-crypt at http://sqlite-crypt.com/download.htm
I don't know whether this one good enough or not. There is a trial version that limit passphrase to 6 characters and store it as plain text in the header. It won't be suitable if you want complete data protection. But for testing, maybe it's worth a try. I don't have a chance to test it yet since I still working on the server side of my project. I'll update it when I've test it in the future.
EDIT: Ok. I've test the trial version and it's worked well with modificated version of SqliteWinRT wrapper on codeplex. Note that the trial is limited to 5-6 characters of passphrase and this phasephrase is stored in plain text, plus it's in x86 compiled binary, which means it won't work for actual product which need ARM support. You have to purchase the binary to get the production-ready binaries.
I am writing a program for linux in C to extract the wpa/wep key from a windows registry hive.
Initially I was hoping to use wine's CryptUnprotectData function, but I realise now that wine uses a different algorithm and just mimics window's version. I also realise that only the user that encrypted the data can decrypt it.
I am using wzcook from the aircrack-ng suite as a guide.
http://tools.assembla.com/b6stFY7MOr2QtlaaeP0Qfc/browser/Windows/wzcook/wzcook.c
I've extracted the data from a hive without problems but I'm kind of stuck on how to decrypt the key.
If the key is encrypted in the registry for one user, then how does another user decrypt that data? I'm sure they don't have to retype the psk / passphrase if it has been already stored?
Is the data definitely encrypted even when offline (accessing the hive from linux)?
Any other solution for solving this, other than reverse engineering Microsoft's algorithm? :D
Any help / advise you can give is really appreciated,
Thanks.
As I understand it DPAPI uses a hash calculated from the user's actual password, so you would need both an open-source implementation of the Crypt functions and an explicit password from the user.
There has been some work on implementing CryptUnprotectData recently: DPAPIck
For Windows.Forms, I have an application that needs to get a unique install id for each install from my server, and then persist this data so once registered, the install ID is included on all communications back to the server. The application is in occasional contact with the server.
How can I persist this data on the client in a way that is not easily tampered with?
First, you should note that if the data is on the local file system and your application can read and write it, it will always be possible for a determined user to tamper with it... perhaps not easy, but possible nonetheless.
That said, there are a number of options you could consider, including (but not limited to) :
encrypting the data with a key defined in your assembly : pretty safe is the user has no programming skills, but an advanced user could disassemble your app to find the key. Obfuscation could make it harder to extract the key, but not impossible.
using an isolated storage : I'm not sure whether the data is encrypted or not, but at least it's not easily found, hidden in a deep folder hierarchy... Not so safe is the user knows where to look, however
writing the data in a binary format, which makes it harder to read or modify for a non expert user
using a piece of native code to encrypt the data : similar to the first option, but a native DLL is harder to disassemble than a .NET assembly, so it requires more skills to find the key
Again, all these approches are not bulletproof : as long as your program can access the data, an advanced user could always reverse engineer it and do the same...
You could save the data in the windows registry. You'll use the [HKCU\Software\YourAppName] hive key if it's a per-user setting, or [HKLM\Software\YourCompany] if it's a global setting. However, the data would need to be encrypted, because its trivila to get the values in these keys
Ok, I know this has been asked a thousand times before, but no conclusive solution has been derived...so here is another really silly question!
I have a Winforms 3.5 app and using LINQ to SQL, hence the Connection string is ALWAYS stored in "app.config" by default (and VS2008 will not accept any other way of storing this - Ive even tried overwriting this in the IDE-generated code). That being said, any person with a bit of computer know-how needs to just browse to the install directory of the app, look for the [appname].exe.config file, and open it to reveal the supposed secure username/password for accessing the database. Even if you opt to encrypt/decrypt this section, it is only done when the app is running - so Im assuming that when the app is closed, the connString section reverts to plain text....that can be once again read. (....how do you win?!?!)
Keeping in mind that Im using LINQ and generating a LinqDataContext, what I would like to know is this:
Can the connstring not be stored elsewhere besides the app.config? (perhaps like in a user setting file that can be modified after installation)
If the above is possible, can I not store a pre-encrypted connstring into the config file (that certainly wont be able to be read) and in my application, opt only to decrypt the connstring whenever the connection is opened?
If none of this is possible, I may have to revert to using traditional ADO.NET (seeing that I already have hundreds of stored procs for all CRUD operations) - at least that way I can have control over how and where the connstring is stored.
BTW - sorry if this is a juvenile/confusing question to ask, and if you feel that Im wrong with anything Im explaining please let me know.
Much thanks!
You can store a pre-encrypted conn-string in the app.config, but i guess it will be useless since, your application sooner or later will decrypt the conn-string (And the plain text will be availiable to anyone interested!)
So you should obtain an SSL cerificate too... (Not a self signed one)