Develop Reference

Best approach for storing sensitive data without using a database or an .xml file

I'm developing an application in c# which uses an .xml file to obtain information that is used inside the logic of the app. When a new version of the app is launched, a setup is created (using Inno Setup Compiler) and after successfully installing the setup, the .xml is placed inside the setup files of the app. This .xml file contains about 200 objects with 4 properties each of sensitive data.
I got asked to launch a customer version of the app, and a requirement for this version is to remove the access or manipulation of this .xml file, since it contains sensitive data that the customer should not be able to see or manipulate.
My senior engineer told me to simply implement the information inside a list in the source code, so that the use of the .xml file is removed and the customer can't manipulate this info once installed the app, as it would be hidden inside the source code, but this seems really inefficient for me and i would need to change a lot of logic about the use of the .xml file inside the app for this to work.
Is there a way to create a setup of the app and hiding this file in the setup files so it cant be manipulated by the customer?
If there isn't, what approach could you suggest me to do? Or do i have no options but to do this the hard way?

If you want to make it harder for the end user to modify the information, while still keeping a separate configuration file that won't require code change of the application itself, you can sign the file and have the application verify the signature.
A simplest way is to to calculate a hash of a file and "secret" value. Of course it is hardly tamperproof. But in the end, there's no tamperproof way to prevent a user from manipulating data on his/hers own computer. It's only about how hard you want to make it.
A better way would be to use a proper certificate for the signature. The application will know only the public key and will use it to verify a signature created with a private key, which will never leave the development team.

From a theoretical standpoint, if your program can get hold of the data, then a user with full control of the computer on which the program is running can also get hold of the data if they try hard enough. That means you can make it difficult for them, but you can't make it impossible. So if that's what you're trying to achieve, you need to be quite clear about the limitations of the approach.
How difficult should you make it? Well, if it's purely a commercial risk, you should make it hard enough that the cost of getting the data is greater than the benefit. If the risk can't be measured in that way, for example if there are legal requirements for you to protect the data, then that isn't going to be good enough.
For some situations, it's probably enough to encrypt the XML file, and bury the decryption key deep within the logic of a compiled program written in a language that doesn't allow easy decompilation. That's likely to be better than simply burying the XML data within the compiled program, which is what your senior engineer is suggesting. But her suggestion may be OK too. It really shouldn't be too difficult to change the program logic from reading an external file to reading a string constant within the program.

Hash of source codes at compile time in C#

Having a server that other devs use, I currently log the version of the dll they use. I do that by having the client that use Reflection to retrieve its version:
Assembly.GetEntryAssembly().GetName().Version.ToString();
It's nice, but since it come from dev that uses TFS and do themself the build, I can not see if they have the latest version of the sources. Is there a trick, like a compilation tag, that would easily allow a hash of the generating source code?
Note: I have try to send the MD5 of the dll (using assembly.Location), but it is useless since the hash value changes between 2 compilations (I suppose there is some compilation timestamp inside the generated dll).

This is most collaboraton issue then a coding.
In the moment that you find out that the version is old one.notify them about it.
If the real version is not old one, that means that developers before making buold did not increment the version ID, which is mistake.
In other words, ordanize it among people, and not relly on these kind of tools (if there is any). You trying to create a complicated tool, that will help you avoid mistakes, but humans will find a way to make them again.
So it's better to create solid relation structure among you, imo.

Create a tool on pre build event to hash/last-write-time your code files.
Write the result to a cs file or a embedded resource file.
The result file must exclude in above action.
For prevent skip build (up-to-date) feature not work,Compare the file before write.
And if youre opening the file in IDE will get a prompt `changed from out side' when build.

Seem there is no easy way to do it.

How to disable an exe file after first installation?

Does anybody know the solution for this? I create an exe file of my software. After first installation I have to disable the exe, so it cannot be run again because when someone purchases the software from me they can install it only once.

To do this you'll need to store something somewhere, that something could be:
A file
A registry entry
A call to a web service you own that stores a unique identifier for the machine, and is checked on subsequent installation attempts (Note: If you choose this method you must be clear and up-front with your users that it's what you're doing).
Bear in mind that a determined user will be able to circumvent file and registry methods and also quite possibly the web service method. The former two by using something such as Process Monitor to identify the files/registry entries you're writing to and clear them. For the latter, by using something like Fiddler to identify the web service calls you're making and replacing the responses with ones that allow them to bypass your protection.
Remember, ultimately the user can disassemble your code and remove the protection mechanisms you've put in place, so don't rely on them being 100% un-breakable

Forget it, mate. It's software - you absolutely cannot enforce something like that because the user has complete control over the environment where the binary runs, including reverse engineering, virtualization, backups etc. etc. And the ones who you want to foil are precisely the ones who will go to any length to thwart any protection measure you could invent.
No, the only thing that works is to force an online connection and register, on your system, the fact that a particular binary was installed once, then forbid it the next time. That requires you to make each installer different and have a cryptographically strong key generator, and it's still susceptible to replay attacks - but it's the only thing that is not useless by definition.
(Well, either that, or make your software so insanely great that people will fall in love you and want to give you the money. That solution is probably even harder.)

You could store the installation path in the registry or some secret location and have your .exe check that if it has started from a location different than the one stored, to simply exit, as you probably don't want to tell the user what you are doing.

How to limit use to only one computer per installation?

I'm about to release my application which is built in C# VS2008 to my customer, and I want to prevent copy abuse post deployment, since it's easy to copy installation files to another machine and use the application. I want to limit usage to only one computer per installation.

See this question for some products that will help you do this...
https://stackoverflow.com/questions/118031/best-activation-key-software-for-net-application
My favorite for now is IntelliLock. Decent price, supports ASP.Net, and has been around for a while.
Locking to a machine ID is not a trivial thing to do manually. So I would us a license software package. Even if you're just trying to deter casual copying, you have to consider machine components change, and people get new computers.
But if you really want to do this manually see CPU serial number and http://www.vcskicks.com/hardware_id.php. But note even the CPU Serial is not a fool proof method as it quite often is disabled.

Use hashing to generate an unlock key. The idea is to gather some data which is fixed on the target machine but also unique. Examples are the name of the machine, id of network card, ... Generate a hash from these values and let the user send this data to you. Generated a new hash from this value and a secret key (only known by you) and send it back to the user. Now the user has to enter this key to unlock your software.

Use an "activation" scheme, like Microsoft does with Windows. Each installation must authenticate itself against a server somewhere using a key. If a key is used more than once, prompt the user to call and talk to a real person.

Just make a pre-screen to enter login\pwd

How can I secure an "enabled functions" license file for my program?

My Application can perform 5 business functions. I now have a requirement to build this into the licensing model for the application.
My idea is to ship a "keyfile" with the application. The file should contain some encrypted data about which functions are enabled in the application and which are not. I want it semi hack proof too, so that not just any idiot can figure out the logic and "crack" it.
The decrypted version of this file should contain for example:
BUSINESS FUNCTION 1 = ENABLED
BUSINESS FUNCTION 2 = DISABLED.... etc
Please can you give me some ideas on how to do this?

While it could definitely be done using Rijndael, you could also try an asymmetric approach to the problem. Require the application to decrypt the small settings file on start up using a public key and only send them new configuration files encrypted using the private key.
Depending on the size of your configuration file, this will cause a performance hit on startup compared to the Rijndael algorithm, but even if the client decompiles the program and gets your public key its not going to matter in regards to the config file since they won't have the private key to make a new one.
Of course none of this considers the especially rogue client who decompiles your program and removes all the checking whatsoever ... but chances are this client won't pay for your product no matter what you do thus putting you in a position of diminishing returns and a whole new question altogether.

Probably the easiest secure solution is to actually use online activation of the product. The client would install your product, enter his key (or other purchase identification -- if you purchase online this could all be integrated, if you are selling a box, the key is more convenient).
You then use this identification to determine what features are available and send back an encrypted "keyfile" (as you term it), but also a custom key (it can be randomly generated, both the key and key file would be stored on your server -- associated with that identification).
You then need to make sure the key file doesn't work on other computers, you can do this by having the computer send back it's machine ID and use that as added salt.

I've been pondering using custom built assemblies for the purpose of application licensing. The key file approach is inherently flawed. Effectively, it's a bunch of flags saying "Feature X is enabled, Feature Y is not". Even if we encrypt it, the application will have all the functionality built in - along with the method to decrypt the file. Any determined hacker is unlikely to find it terribly hard to break this protection (though it may be enough to keep the honest people honest, which is really what we want).
Let's assume this approach of encrypted "Yay/Nay" feature flags is not enough. What would be better is to actually not ship the restricted functionality at all. Using dynamic assembly loading, we can easily put just one or two core functions from each restricted feature into another assembly and pull them in when needed. These extra "enablement" assemblies become the keyfiles. For maximum security, you can sign them with your private key, and not load them unless they're well signed.
Moreover, for each customer, your build and licensing process could include some hard to find customer specific data, that effectively ties each enablement assembly to that customer. If they choose to distribute them, you can track them back easily.
The advantage of this approach over simple Yay/Nay key files is that the application itself does not include the functionality of the restricted modes. It cannot be hacked without at least a strong idea of what these extra assemblies do - if the hacker removes their loading (as they would remove the keyfile), the code just can't function.
Disadvantages of this approach include patch release, which is somewhat mitigated by having the code in the keyfile assemblies be simple and compact (yet critical). Custom construction of an assembly for each customer may be tricky, depending on your distribution scenario.

You could achieve this fairly easily using Rijndael, however the problem is the fact that the code will contain your Key in your current design. This basically means someone will disassemble your code to find the key and boom, goodbye protection. You could slow this process by also obfuscating your code, but again, if they want to get it, they will get it.
However, this aside, to answer your question, this code should work for you:
http://www.dotnetspark.com/kb/810-encryptdecrypt-text-files-using-rijndael.aspx

I find Perforce-style protection scheme easiest to implement and use, while at the same time being quite hack-proof. The technique uses a plain text file with a validation signature attached at the last line. For example:
----(file begin)
key1: value1
key2: value2
expires: 2010-09-25
...
keyN: valueN
checksum: (base64-encoded blob)
---- (file end)
You would choose an assymetric (public/private key) encryption algorithm + hashing algorithm of your choice. Generate your reference public/private key pair. Include the public key in your program. Then write a small utility program that will take an unsigned settings file and sign it - compute the digital signature for the contents of the file (read settings file, compute hash, encrypt this hash using private key) and attach it (e.g. base64-encoded) as "checksum" in the last line.
Now when your program loads the settings file, you would read the embedded public key and validate the digital signature (read file contents, strip the last line, compute hash; compare this value against checksum from last line base64 decoded and run through the assymetric decryption using embedded public key). If the validation succeeds, you know the settings file has not been tampered with.
I find the advantages to be that the settings are in plain text (so for example the customer can see when the license expires or what features they paid for), however changing even a single character in the file with result in the digital signature check failing. Also, keep in mind that you are now not shipping any private knowledge with your program. Yes, the hackers can reverse-engineer your program, but they will only find the public key. To be able to sign an altered settings file, they will have to find the private key. Good luck doing that unless you're a three-letter agency... :-).

Use any 'Cryptography' method to implement this.
Just check out the namespace 'System.Security.Cryptography'
The above namespace providing many encryption and decryption functions to protect secret data.
You have another method to implement this using registry.
You can store data in windows registry.
Better to encrypt data before store into registry.

ROT-13!
Edit:
ROT-13 is a simple substitution cipher in which each letter is substituted by the letter 13 letters before it in the alphabet. (NOTE: alternatively, you can use the ascii-value 13 less than the given char to support more than [ A-Z0-9]).
For more info see wikipedia.