Suppose I have logged into an application which is running from IIS. Now I haven't logged out, but closed the browser. And when I'm accessing the application again, it defaults to the login page. How does IIS recognize that it is a new request and redirects the user to the login page?
I have another question. Suppose if I'm not closing the browser, which I used when I logged in. I'm opening the new browser to request a page from same application. IIS recognizes that it's a new request to the application and redirects the user to login page. Why does it not use the existing session or cookies which the first browser uses?
Please don't get irritated of my continuous questions... I am having huge confusion.
We say HTTP is a stateless protocol. Once the page is requested I have logged in. And the HTTP protocol connection will be terminated between IIS and browser, right? Then I am navigating to other pages in that logged in application. Now IIS recognises the user has logged in on this browser. But when I open a new browser and request that application, how does IIS recognises it is a new request? Since the HTTP protocol is disconnected, how does it work in the first case?
As you've correctly said, HTTP itself is stateless, and each request is technically separate from every other. Sessions, as used by web sites, are a workaround for that. What happens, normally, is that the server stores whatever info it cares to maintain between requests (like the logged-in user's username and/or ID, for example), and assigns that information an ID (called a "session ID"). It then tells the browser that session ID, in such a way that the browser can hand the ID back when it's time to make another request. If the browser plays its part and provides the session ID, then the stored information can be retrieved, updated, etc with each request, providing some degree of state even over a stateless protocol.
Sessions are usually implemented using cookies. That is, the server hands the browser a cookie with the session ID, and the browser hands back that same cookie with each request until the cookie expires or is otherwise forgotten. Some cookies (so-called "session cookies") aren't saved, and are forgotten when the browser is closed. A freshly opened browser doesn't have any session cookies to pass, so if the server uses session cookies to do sessions (which it should), it will consider the user not yet logged in and bounce them to the login page if they need to be logged in.
Session cookies will usually be shared between tabs in the same browser, and will sometimes even be shared by windows opened by "File > New Window" from an already running browser, because both of those cases will typically just be a part of that browser. But if you start the browser from the Start menu, or however your OS lets you start a program, it's a whole other process -- and session cookies are rarely shared between processes.
The server typically also only remembers sessions on its end for a limited time (anywhere from seconds to years, depending on the server and/or site settings) after each request that uses the session. If the browser passes a cookie that corresponds to a session the server no longer remembers, it'll act as if there's no session at all. Which, in cases where you have to log in, will again bounce to the login page.
There are cookies that are passed always no matter are you logged or not. They are mapped to session in IIS.
Check out the following articles. They might be helpful.
IIS Dropping Sessions
Session Management in ASP.NET
Related
I am using an MVC app to manage authentication. The issue I have is with chrome because it never actually kills the session because it runs in the background after it closes by default. I do not want to enforce all the end users to change this setting because then it will kill hangouts etc.. So I am wondering if I can use any standard web.config setting to handle this or do I need to make an ajax polling interval to keep updating the cookie expiration?
The ASP.NET session consists of 2 components the client-side http session cookie and the server's session storage provider. Suppose user has SessionId 1. If you delete session 1 from the server, the user returns and a new session is created with SessionId 1. If the user deletes the session cookie, the server keeps running session 1.
What you're asking for is generally not possible. There's no way you can force a user to send a request to your server when they are exiting your site or the browser. There is the javascript beforeUnload event which in some situations would allow you to send a request to /sign-off in some situations. The obvious limiter is no network access = no message.
The standard resolution for orphaned sessions is for session scavenging policy to clean them up. Some developers choose to use persistent storage to eliminate scavenging altogether such that a shopping cart would never disappear.
The only reasonable solution (which is still overkill) that would reach your goals. You use SingalR for a persistent connection of the user to the server and you ping them from the server. If the connection fails to respond you abandon the session. This will be a fragile process and if you don't make very very sure the user is disconnected you will have lots of support calls from users wondering why they are continuously logged out while browsing your site on cell phone.
If the session is stored in proc
The user logs in, closes the browser directly and reopens it after an hour. Would he need to log in again?
If the web application uses cookies and cookies are enabled on the users browser...
If the web application uses persistent cookies and cookies are enabled on the users browser...
If the web application uses cookies and cookies are disabled on the users browser...
If the web application uses persistent cookies and cookies are disabled on the users browser...
If session is stored in state server and situations are the same, then what would happen?
When a session is created (assuming it is a normal session), a cookie is sent to the browser that looks something like this:
SESSION_ID=437b930db84b8079c2dd804a71936b5f
Sessions can be used without cookies if the session identifier (in the example above, 437b930db84b8079c2dd804a71936b5f) is passed around as request parameter instead of a cookie; however, this is rather uncommon and it is generally considered bad practice.
All session information is stored server-side, and the session identifier is used behind-the-scenes to decide which set of information should be recalled (if any) for each request. Thus we get to your questions.
If the web application uses cookies and cookies are enabled on the users browser...
If the web app uses cookies and cookies are enabled on the browser, then there should not be a problem. With a standard session implementation, the cookies will be non-persistent, though, so the user will need to login again if he/she completely closes all instances of the browser.
If the web application uses persistent cookies and cookies are enabled on the users browser...
If the session-id is stored in a persistent cookie and user's browser respects that by persisting the session identifier cookie to the disk, then the session identifier will be sent even if the browser is fully closed and restarted. However, please be aware that most web frameworks have a garbage-collector-like system that deletes data for sessions that have showed any activity over a certain amount of time. So, for example, let's say my website requires activity at least once every 4 hours to keep a session active. If I login, receive a persistent cookie with my session ID, close my browser, and come back 5 hours later, then I will need to login again because my session information would have been cleared from the server even though my session ID cookie was persistent.
If the web application uses cookies and cookies are disabled on the users browser...
Bad news bears. You will either need to find a way to use a cookieless session (passing an identifier as a parameter for each request), or you will need to ask the user to enable cookies. There is no way around this.
If the web application uses persistent cookies and cookies are disabled on the users browser...
Same situation as #3. If the user has cookies disabled, you are out of luck. Either they need to enable cookies (at least for your site), or you need to find another way to pass around information between requests.
Session is stored in server memory (unless a state server or persistant store is used) but relies on a cookie to identify the session. If cookies aren't available then session won't work since there is no way to identify the user. Cookieless sessions can be used to get around this. Cookieless sessions aren't recommended as they can be hijacked with the session identifier in the url.
If an expiration isn't set on the cookie then it will be lost once the user closes all browser instances (they share memory) and not just the one visited through the website.
If the user has cookies disabled, then cookies aren't available for use by the application. People aren't as worried about cookies now as they were in the late 90's (lots of "security" people raised warnings that cookies could be used to store all sorts of things on your computer, even viruses).
What is the common(best practice) way of allowing a user to sign in from a Windows desktop application. Some examples of what I mean are Dropbox or Google Picasa. You sign in with your credentials and then the software is permanently signed in.
I assume the communication takes place over HTTPS. Does the client store the credentials to be sent with requests or is there some sort of token generated? Can anyone point me to some resources on how this should be handled?
Logging into a website normally creates a session at the server. The server then has to identify subsequent requests by the session. Typically, there is one of the following two solutions applied:
Session cookie, storing a session identifier
URL rewriting, where the session identifier is appended to every link in the html source
Which approach is taken is site dependent, so if you are writing a general 'for all sites' client, you might have to implement both.
In the former case, your application will have to handle the session cookies, in the second case, your application has either nothing to do - if it caches the html response - or will have to emulate the url rewriting itself.
In both cases be aware that the session will expire at server side after a certain period without any activity, so you might be required to generate such.
In my previous understanding, for a web site, only login user of a web site (no matter what login/authentication approach is used) could have cookie as persistent identifier, so that if the user close the browser, open browser again to go to the same web site, the web site could remember the user.
But I learned recently that it seems for non-login user, there could still be a cookie associated with the user (after the user close browser, and then open the browser again to go to the same web site, the web site could remember the user), and it is called browser cookie? Is that true?
If it is true, who is responsible to set the browser cookie? i.e. need some coding/config at web server side, client browser configuration (without coding from server side), or both? How could web server access such cookie? Appreciate if any code samples.
thanks in advance,
George
Whether you actually "log in" or not is irrelevant to what cookies are stored.
If the browser requests a page, and the server includes a Set-Cookie response header, then the browser will store the value of that cookie in a local cache and every time it requests a page from the same server, it sends the value of the cookie back as well (in the Cookie request header).
It just so happens that when you "log in" to a website, the website will usually use the Set-Cookie header to tell the browser to store a value that indicates that you're already logged in (and your user-id and some other security-related stuff). But there's nothing stopping the web server from using Set-Cookie at any other time.
I'm working on an ASP.NET application where our users authenticate using client certificates over HTTPS. Our users are only using IE7.
Once a client certificate has been used to authenticate successfully it stays in the browser's SSL cache until the process is closed or the user manually clears the SSL cache. We want to be able to clear the SSL cache whenever a user logs out, or their session expires, to improve the security of the system.
Our clients already use smartcards to access the system, which unload certificates automatically when the card is removed from the client computer, but this does not clear the browser cache at all, leaving a potential avenue of attack from another user who had access to the same machine as the genuine user.
I've found out how to do the actual cache clearing from JavaScript:
document.execCommand("ClearAuthenticationCache");
which works perfectly when a user explicitly logs out, as we can execute the script on the client before allowing the user to log in again.
NOTE: IE7 only lets the cache be cleared programmatically when HTTP Keep-Alives are disabled on the web server.
Here's the tricky bit - if a client's session expires, I don't know of any way to handle this in the browser before the user tries to login again. I can't clear the state when they get to the login page, because I need the state cleared and a new certificate chosen before the page executes on the server.
Any ideas? Apologies for length of question, but background is important for this one.
Never mind, I came up with a good solution:
When the user successfully logs in, we create an additional session cookie that doesn't expire until the browser is closed.
If the user comes back to the login page later and the request is unauthenticated, we check for the existence of the session cookie - if it exists, we know that the user has previously had a session, so we explicitly log them out, exactly as we do for the user-initiated logout. If the session cookie doesn't exist then we attempt to automatically log the user in using their certificate.
The custom session cookie is deleted for each explicit log out, and re-populated for each successful login.
This gives us the best experience for the user, and guarantees that a certificate will be cached only as long as a session is still valid (15 minutes, sliding). Also, the session cookie cannot be removed by the user so there is no way to bypass this behaviour. They can't use the site without accepting session cookies either.