Take the 2-minute tour ×
Programmers Stack Exchange is a question and answer site for professional programmers interested in conceptual questions about software development. It's 100% free, no registration required.

I need to build a web service API for our mobile app to interact with our server & database (in ASP.Net MVC 4, but that's hardly relevant). Wile most actions do not need users to be registered with our service, we would like to restrict access only to users of our app.

What are the methods to make sure that calls made from somewhere else (e.g. someone wanting all our data, or building an unofficial app) are rejected?

My initial idea is for the device to ask the server for a token, which is randomly generated and sent as is. All other methods will check that the request headers contain a specific one that will be md5 hash of the salted token. The server knows the token and salt and is able to compute the hash and compares it with the one sent. In addition, the token as a limited lifespan, and the device must get another one every other hour.

That seems quite easy to implement and although probably not 100% proof, that seems to be a good idea. What do you think?

share|improve this question
add comment

1 Answer

up vote 7 down vote accepted

As soon as you release the application, it might be reverse engineered. This means that there is nothing you can do to be 100% protected if the same application (same binaries, same settings) is distributed to all of your users.

If you can customize the application for every user, then you have a chance to maybe not forbid some other application to use your API, but at least limit this application by number of requests it can do to the API.

Imagine the following schema:

  1. Client connects and sends its unique identifier (one identifier by user).
  2. Server replies by sending a challenge encrypted with a public key. This public key is associated with the unique identifier previously sent.
  3. Client solves the challenge by decrypting the data using a private key and sends the decrypted secret in plain back to the server.
  4. Server verifies that the submitted secret corresponds to the one originally generated.

The developer who will hack your application and successfully obtain the private key would be able to use your API from his own application, but he will be identifier as himself for your server.

If the same user can do 10 000 requests to your API per day and in average, an active user makes 2 000 requests per day, it means that this developer would be able to use his application himself and maybe give it to his friends, but he wouldn't be able to, say, sell it to thousands of people, just because it will work only for a few minutes in the morning.

While this helps, it's not 100% proof neither. What if the hacker finds a way to extract the private key from your app when his own app is installed on the device?


Side note which doesn't answer your question, but still might be useful: don't think about an API as a tool for your primary product (mobile application). Think about it as a first class product itself, a product which may be paid. The same model is used for years by Amazon and Google, it starts to be actively used by Microsoft with Azure, etc.

As soon as you consider the API not as a secondary tool reduced to slavery for your shiny new mobile apps, but the actual product, at the same level as any application the user actually sees, you start thinking less about how to protect the API against the usage by other apps, and more about the monetization of the API itself. Such API can be used by your apps which are its customers, or any other apps, freely developed by anyone. This has several benefits:

  • Making an API so that it would be used only by your applications is difficult and expensive. This time and money can be used for something more useful.

  • Opening your API to the public can have a great benefit for both you and the world. Imagine you're a great architect and a great developer, so you've created an astonishingly great API, but your visual designer skills suck and you really don't understand anything about interaction design, etc. If you hide your API, the only thing people will know is that you've created a mobile application which is unusable and ugly. If your API is public, other developers will be attracted by its quality and write great applications for it, bringing you lots of money.

  • You never imagine how other people may use your APIs. This is what happened with Kinect. Originally, Microsoft created Kinect for games. When Microsoft opened the API to the public, they never imagined it would be used a few years later by scientific applications, health sector, etc. It's similar for web APIs: more developers are using it, more widespread the ideas would be.

share|improve this answer
    
Very interesting answer. We probably need to think how we can make the API public. Also for reference, the target market in France, where reverse engineering and disassembling is illegal, so we should preoccupy too much with this. –  Antoine Apr 29 '13 at 7:00
2  
@Antoine: Revere engineering would be illegal in any country as soon as the license prohibits it. But the fact that it is illegal doesn't mean that nobody will revere engineer your app. –  MainMa Apr 29 '13 at 7:24
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.