Jersey Web Services Security Question
I am building a public facing REST application for clients to access. In trying to develop security for it, I looked at how AWS handles their REST authentication using an access key and secret key to sign the message before sending to the AWS server. So I coded an example with Jersey, JSP, and javascript (jsonp) to test, and it seems to work ok.
1.) On page load the JSP takes the client's access key and date/time (converted to EPOCH) and writes out to the page. It also takes those two pieces of information and combines them with the client's unique secret key via HMAC-SHA1.
2.) When I execute the JQuery ajax request, I pass in all three pieces of information.
- user access key
- date/time (EPOCH)
- signed message
3.) On the Jersey side I take the access key and date/time and check a database for the users secret key. Then I perform my own HMAC-SHA1 s开发者_JAVA百科ignature and compare that to the signed message from the user submitted signed message. If they match, then allow them to use our web service. I also check the date/time and only accept messages signed within the last 15 minutes.
My question is, is this secure enough?
This isn't a complete answer, but I don't have the points to just comment quite yet. I would add to your security a salt on the server side. Please see this article for a good discussion on hashing using Master Keys and salts.
精彩评论