Basic info
I have a webserver with all that fancy stuff.
It runs on my half-potato PC, so I don't have any free 64TB of SSD storage.
I have to care about scalability, because one day I might have a 1000 users.
Wow, such users, very a lot!
I don't want any passwords to be stolen if user will lose his device. So I decided to use sessions.
I don't want to store all of the user sessions, because that can be possibly a 100 of devices. So 1000 * 100 is a lot of session cookies.
I thought about a way not to store, but to validate sessions to check if request is from the owner of the account.
Structure
Global
// Secret 128 bytes array.
const ByteArray serverMagic;
// Generates a SHA512 hash based on data.
function hash(ByteArray data);
// Encodes array of bytes into base64 string.
function base64(ByteArray data);
// Returns current time on the server.
function getCurrentDate();
// Returns userMagic based on the user ID.
function getUserMagic(Integer uid);
// Returns true if user entered wrong password.
function isLoginInvalid(Integer uid, String password);
function Object.toShortString(); // Encodes an object into a short ASCII string.
For each user
const Integer uid; // User ID.
var String password; // Secret code that gives you all of that damn power.
var ByteArray userMagic // Secret 128 bytes array.
Logic
If user wants to log in
he will send:
Session expiration date (so cookie won't be valid after a desired amount of time)
User ID
Password (or hash of the password, it doesn't really matter)
function logIn(Date expire, Integer uid, String password) {
if(isLoginInvalid(uid, password)) then
return error;
String encExpire = expire.toShortString();
String encUID = uid.toShortString();
ByteArray userMagic = getUserMagic(uid);
ByteArray food = encExpire + '_' + serverMagic + userMagic + '_' + encUID;
ByteArray hash = hash(food);
String session = encExpire + '_' + encUID + '_' + base64(hash);
return session;
}
Client will save that session cookie.
If user will make some action
he will send a session cookie to the server.
// After parsing session cookie
function request(Date sessionExpire, Integer uid, ByteArray sessionHash) {
if(sessionExpire < getCurrentDate()) then
return error;
String encExpire = sessionExpire.toShortString();
String encUID = uid.toShortString();
ByteArray userMagic = getUserMagic(uid);
ByteArray food = encExpire + '_' + serverMagic + userMagic + '_' + encUID;
ByteArray hash = hash(food);
if(hash == sessionHash) then
return success;
else
return error;
}
If user will want to invalidate all sessions
due to device loss or password change server can just generate a new userMagic, that will make all of the previous session cookies invalid.
Questions
Is it a good way to store not to store sessions?
Is there a flaw in my system?
Can it be cracked?
And how can it be cracked?
How can it be abused?
How can I improve my system?
Revised answer:
The problem with solutions like this is that they are often vulnerable to very subtle bugs, such as length extension attacks. Play it safe and use HMAC instead.
FYI .Net has some construction that is not too different than this, but uses HMAC and no user magic.
Previous, retracted answer:
Any user can compute:
String session = encExpire + '_' + encUID + '_' + base64(hash);
Which means any user can create arbitrary sessions. The hacker is going to do this with encUID set to the administrator uid, which may 1 or 0 or something like that.
I could potentially suggest fixing this bug using HMAC with a secret key, but you really should not be implementing this yourself. Find a widely used and well studied library that does session management for you rather than trying to re-invent the wheel.
I'm trying to login user with jaas using Scrypt. I was using sha before and I could just login using:
request.login(username, "password");
and jaas would make the check in the background. But since the method to check if 2 passwords match with scrypt is a check that returns a boolean I had to change the code a bit. I think my code is correct/safe but I prefer being safe than sorry:
// here I get user credential from the db
Usercredential uc = us.getUsercred(username);
boolean matched = false;
if (null != uc) {
// Making the scrypt check to see if password entered by user match db pass matched
matched = SCryptUtil.check(password + uc.getSalt(), uc.getPassword());
}
if (matched) {
// if pass matched I login with the db password value
try {
request.login(username, uc.getPassword());
}...
I'm worried about the fact I'm logging in with the DB value and not the typed value. I make a check to see if they match prior to that though so there should be no problem.
After following the guide for creating a custom identity provider for azure mobile services I can easily generate the appropriate tokens. The code is pretty simple and looks like this:
var userAuth = {
user: { userId : userId },
token: zumoJwt(expiry, aud, userId, masterKey)
}
response.send(200, userAuth);
The definitions for the parameters and code for zumoJwt are located at the link. Azure automatically decodes the token and populates the user on the request object which is what I'd like to simulate.
Basically I'd like to to decrypt the token on the serverside via Node (not .net).
What I ended up doing to validate the token is the following (boiled down). This seems to be about what the azure mobile services is doing on routes that require authorization.
var jws = require('jsw'); // https://github.com/brianloveswords/node-jws
function userAuth() {
var token = ''; // get token as header or whatever
var key = crypto.createHash('sha256').update(global.masterKey + "JWTSig").digest('binary');
if (!jws.verify(token,key)) {
// invalid token logic
} else {
var decode = jws.decode(token)
req.user = {
userId: decode.payload.uid.split(';')[0].split('::')[0]
};
next();
}
}
app.use(authChecker);
The tokens aren't really encrypted - they're just signed. The tokens have the JWT format (line breaks added for clarity):
<header>, base64-encoded
"."
<envelope>, base64-encoded
"."
<signature>, base64-encoded
If you want to decode (not decrypt) the token in node, you can split the value in the . character, take the first two members, base64-decode them (var buffer = new Buffer(part, 'base64')), and convert the buffer to string (buffer.toString('utf-8')).
If you want to validate the token, just follow the same steps you need to re-sign the first two parts of the token (header + '.' + envelope) with the master key, in the same way that the token was created, and compare it with the signature you received on the original token.
I have a web app being built in express.js with a postgresql db.
I was wondering about how to implement the security, but everyone uses something different ( i guess thats a good thing? ).
Different modules different authentication sequences etc.
What I have at the moment:
1) User form post to for example /login
2) app routes to specific route
3) in route I try the following
var localconstring = "postgres://" + usr + ":" + pass + "#ip:port/db";
var client = new pg.Client(localconstring);
client.on('drain', client.end.bind(client));
client.connect(function (err, client, done) {
The database uses md5 so the pass is already protected by the db.
What should really happen?
Should I salt and hash the username and password and then save the salted/hashed credentials alongside the salt and then use the md5 of the db also?
If so which module?
Should I be logging in like that or try to do a select * from pg_roles/users ??
Thanks a lot!
(regarding the salt and hash if possible some detailed examples as I am pretty knew with authentication security)
Forgot to mention. cookies..
After the authentication I set the following cookies:
res.cookie('user', req.body.lguser.username, { signed: true })
res.cookie('watcher', o, { signed: true })
And look em up afterwards
req.signedCookies.user !== undefined
Is the signed attribute secure?
You should generate a key. This key should be saved on a cookie and on the database. Then when the user makes a petition, you can get the key on the cookie and search the user on the database.
There are libraries that help you on this, take a look at Passportjs:
http://passportjs.org/
First of all md5 is NOT seure anymore, so I would recommend you using 'sha512'.
A snippet would be something like this:
var crypto = require('crypto');
var salt = crypto.pseudoRandomBytes(32);
crypto.pbkdf2(userPassword,salt,1024,32,function(err,finalPassword){
//on the db you save the salt as a field and the SALTEDPASSWORD !!
//in this case the finalPassword
}
So when the user logs-in you get the user from the db by username and do the following:
//after getting the user from DB recalculate the hash
crypto.pbkdf2(passw,user.salt,1024,32,function(err,corrPass){
if(corrPass.toString() == user.password.toString()) // log in the user
//where user.password is the result from the db query
}
And I do recommend using passport like the other dude said, it simplifies all of the cookie stuff.
Hope it helped !
In my build process, I want to include a timestamp from an RFC-3161-compliant TSA. At run time, the code will verify this timestamp, preferably without the assistance of a third-party library. (This is a .NET application, so I have standard hash and asymmetric cryptography functionality readily at my disposal.)
RFC 3161, with its reliance on ASN.1 and X.690 and whatnot, is not simple to implement, so for now at least, I'm using Bouncy Castle to generate the TimeStampReq (request) and parse the TimeStampResp (response). I just can't quite figure out how to validate the response.
So far, I've figured out how to extract the signature itself, the public cert, the time the timestamp was created, and the message imprint digest and nonce that I sent (for build-time validation). What I can't figure out is how to put this data together to generate the data that was hashed and signed.
Here's a rough idea of what I'm doing and what I'm trying to do. This is test code, so I've taken some shortcuts. I'll have to clean a couple of things up and do them the right way once I get something that works.
Timestamp generation at build time:
// a lot of fully-qualified type names here to make sure it's clear what I'm using
static void WriteTimestampToBuild(){
var dataToTimestamp = Encoding.UTF8.GetBytes("The rain in Spain falls mainly on the plain");
var hashToTimestamp = new System.Security.Cryptography.SHA1Cng().ComputeHash(dataToTimestamp);
var nonce = GetRandomNonce();
var tsr = GetTimestamp(hashToTimestamp, nonce, "http://some.rfc3161-compliant.server");
var tst = tsr.TimeStampToken;
var tsi = tst.TimeStampInfo;
ValidateNonceAndHash(tsi, hashToTimestamp, nonce);
var cms = tst.ToCmsSignedData();
var signer =
cms.GetSignerInfos().GetSigners()
.Cast<Org.BouncyCastle.Cms.SignerInformation>().First();
// TODO: handle multiple signers?
var signature = signer.GetSignature();
var cert =
tst.GetCertificates("Collection").GetMatches(signer.SignerID)
.Cast<Org.BouncyCastle.X509.X509Certificate>().First();
// TODO: handle multiple certs (for one or multiple signers)?
ValidateCert(cert);
var timeString = tsi.TstInfo.GenTime.TimeString;
var time = tsi.GenTime; // not sure which is more useful
// TODO: Do I care about tsi.TstInfo.Accuracy or tsi.GenTimeAccuracy?
var serialNumber = tsi.SerialNumber.ToByteArray(); // do I care?
WriteToBuild(cert.GetEncoded(), signature, timeString/*or time*/, serialNumber);
// TODO: Do I need to store any more values?
}
static Org.BouncyCastle.Math.BigInteger GetRandomNonce(){
var rng = System.Security.Cryptography.RandomNumberGenerator.Create();
var bytes = new byte[10]; // TODO: make it a random length within a range
rng.GetBytes(bytes);
return new Org.BouncyCastle.Math.BigInteger(bytes);
}
static Org.BouncyCastle.Tsp.TimeStampResponse GetTimestamp(byte[] hash, Org.BouncyCastle.Math.BigInteger nonce, string url){
var reqgen = new Org.BouncyCastle.Tsp.TimeStampRequestGenerator();
reqgen.SetCertReq(true);
var tsrequest = reqgen.Generate(Org.BouncyCastle.Tsp.TspAlgorithms.Sha1, hash, nonce);
var data = tsrequest.GetEncoded();
var webreq = WebRequest.CreateHttp(url);
webreq.Method = "POST";
webreq.ContentType = "application/timestamp-query";
webreq.ContentLength = data.Length;
using(var reqStream = webreq.GetRequestStream())
reqStream.Write(data, 0, data.Length);
using(var respStream = webreq.GetResponse().GetResponseStream())
return new Org.BouncyCastle.Tsp.TimeStampResponse(respStream);
}
static void ValidateNonceAndHash(Org.BouncyCastle.Tsp.TimeStampTokenInfo tsi, byte[] hashToTimestamp, Org.BouncyCastle.Math.BigInteger nonce){
if(tsi.Nonce != nonce)
throw new Exception("Nonce doesn't match. Man-in-the-middle attack?");
var messageImprintDigest = tsi.GetMessageImprintDigest();
var hashMismatch =
messageImprintDigest.Length != hashToTimestamp.Length ||
Enumerable.Range(0, messageImprintDigest.Length).Any(i=>
messageImprintDigest[i] != hashToTimestamp[i]
);
if(hashMismatch)
throw new Exception("Message imprint doesn't match. Man-in-the-middle attack?");
}
static void ValidateCert(Org.BouncyCastle.X509.X509Certificate cert){
// not shown, but basic X509Chain validation; throw exception on failure
// TODO: Validate certificate subject and policy
}
static void WriteToBuild(byte[] cert, byte[] signature, string time/*or DateTime time*/, byte[] serialNumber){
// not shown
}
Timestamp verification at run time (client site):
// a lot of fully-qualified type names here to make sure it's clear what I'm using
static void VerifyTimestamp(){
var timestampedData = Encoding.UTF8.GetBytes("The rain in Spain falls mainly on the plain");
var timestampedHash = new System.Security.Cryptography.SHA1Cng().ComputeHash(timestampedData);
byte[] certContents;
byte[] signature;
string time; // or DateTime time
byte[] serialNumber;
GetDataStoredDuringBuild(out certContents, out signature, out time, out serialNumber);
var cert = new System.Security.Cryptography.X509Certificates.X509Certificate2(certContents);
ValidateCert(cert);
var signedData = MagicallyCombineThisStuff(timestampedHash, time, serialNumber);
// TODO: What other stuff do I need to magically combine?
VerifySignature(signedData, signature, cert);
// not shown: Use time from timestamp to validate cert for other signed data
}
static void GetDataStoredDuringBuild(out byte[] certContents, out byte[] signature, out string/*or DateTime*/ time, out byte[] serialNumber){
// not shown
}
static void ValidateCert(System.Security.Cryptography.X509Certificates.X509Certificate2 cert){
// not shown, but basic X509Chain validation; throw exception on failure
}
static byte[] MagicallyCombineThisStuff(byte[] timestampedhash, string/*or DateTime*/ time, byte[] serialNumber){
// HELP!
}
static void VerifySignature(byte[] signedData, byte[] signature, System.Security.Cryptography.X509Certificates.X509Certificate2 cert){
var key = (RSACryptoServiceProvider)cert.PublicKey.Key;
// TODO: Handle DSA keys, too
var okay = key.VerifyData(signedData, CryptoConfig.MapNameToOID("SHA1"), signature);
// TODO: Make sure to use the same hash algorithm as the TSA
if(!okay)
throw new Exception("Timestamp doesn't match! Don't trust this!");
}
As you might guess, where I think I'm stuck is the MagicallyCombineThisStuff function.
I finally figured it out myself. It should come as no surprise, but the answer is nauseatingly complex and indirect.
The missing pieces to the puzzle were in RFC 5652. I didn't really understand the TimeStampResp structure until I read (well, skimmed through) that document.
Let me describe in brief the TimeStampReq and TimeStampResp structures. The interesting fields of the request are:
a "message imprint", which is the hash of the data to be timestamped
the OID of the hash algorithm used to create the message imprint
an optional "nonce", which is a client-chosen identifier used to verify that the response is generated specifically for this request. This is effectively just a salt, used to avoid replay attacks and to detect errors.
The meat of the response is a CMS SignedData structure. Among the fields in this structure are:
the certificate(s) used to sign the response
an EncapsulatedContentInfo member containing a TSTInfo structure. This structure, importantly, contains:
the message imprint that was sent in the request
the nonce that was sent in the request
the time certified by the TSA
a set of SignerInfo structures, with typically just one structure in the set. For each SignerInfo, the interesting fields within the structure are:
a sequence of "signed attributes". The DER-encoded BLOB of this sequence is what is actually signed. Among these attributes are:
the time certified by the TSA (again)
a hash of the DER-encoded BLOB of the TSTInfo structure
an issuer and serial number or subject key identifier that identifies the signer's certificate from the set of certificates found in the SignedData structure
the signature itself
The basic process of validating the timestamp is as follows:
Read the data that was timestamped, and recompute the message imprint using the same hashing algorithm used in the timestamp request.
Read the nonce used in the timestamp request, which must be stored along with the timestamp for this purpose.
Read and parse the TimeStampResp structure.
Verify that the TSTInfo structure contains the correct message imprint and nonce.
From the TimeStampResp, read the certificate(s).
For each SignerInfo:
Find the certificate for that signer (there should be exactly one).
Verify the certificate.
Using that certificate, verify the signer's signature.
Verify that the signed attributes contain the correct hash of the TSTInfo structure
If everything is okay, then we know that all signed attributes are valid, since they're signed, and since those attributes contain a hash of the TSTInfo structure, then we know that's okay, too. We have therefore validated that the timestamped data is unchanged since the time given by the TSA.
Because the signed data is a DER-encoded BLOB (which contains a hash of the different DER-encoded BLOB containing the information the verifier actually cares about), there's no getting around having some sort of library on the client (verifier) that understands X.690 encoding and ASN.1 types. Therefore, I conceded to including Bouncy Castle in the client as well as in the build process, since there's no way I have time to implement those standards myself.
My code to add and verify timestamps is similar to the following:
Timestamp generation at build time:
// a lot of fully-qualified type names here to make sure it's clear what I'm using
static void WriteTimestampToBuild(){
var dataToTimestamp = ... // see OP
var hashToTimestamp = ... // see OP
var nonce = ... // see OP
var tsq = GetTimestampRequest(hashToTimestamp, nonce);
var tsr = GetTimestampResponse(tsq, "http://some.rfc3161-compliant.server");
ValidateTimestamp(tsq, tsr);
WriteToBuild("tsq-hashalg", Encoding.UTF8.GetBytes("SHA1"));
WriteToBuild("nonce", nonce.ToByteArray());
WriteToBuild("timestamp", tsr.GetEncoded());
}
static Org.BouncyCastle.Tsp.TimeStampRequest GetTimestampRequest(byte[] hash, Org.BouncyCastle.Math.BigInteger nonce){
var reqgen = new TimeStampRequestGenerator();
reqgen.SetCertReq(true);
return reqgen.Generate(TspAlgorithms.Sha1/*assumption*/, hash, nonce);
}
static void GetTimestampResponse(Org.BouncyCastle.Tsp.TimeStampRequest tsq, string url){
// similar to OP
}
static void ValidateTimestamp(Org.BouncyCastle.Tsp.TimeStampRequest tsq, Org.BouncyCastle.Tsp.TimeStampResponse tsr){
// same as client code, see below
}
static void WriteToBuild(string key, byte[] value){
// not shown
}
Timestamp verification at run time (client site):
/* Just like in the OP, I've used fully-qualified names here to avoid confusion.
* In my real code, I'm not doing that, for readability's sake.
*/
static DateTime GetTimestamp(){
var timestampedData = ReadFromBuild("timestamped-data");
var hashAlg = Encoding.UTF8.GetString(ReadFromBuild("tsq-hashalg"));
var timestampedHash = System.Security.Cryptography.HashAlgorithm.Create(hashAlg).ComputeHash(timestampedData);
var nonce = new Org.BouncyCastle.Math.BigInteger(ReadFromBuild("nonce"));
var tsq = new Org.BouncyCastle.Tsp.TimeStampRequestGenerator().Generate(System.Security.Cryptography.CryptoConfig.MapNameToOID(hashAlg), timestampedHash, nonce);
var tsr = new Org.BouncyCastle.Tsp.TimeStampResponse(ReadFromBuild("timestamp"));
ValidateTimestamp(tsq, tsr);
// if we got here, the timestamp is okay, so we can trust the time it alleges
return tsr.TimeStampToken.TimeStampInfo.GenTime;
}
static void ValidateTimestamp(Org.BouncyCastle.Tsp.TimeStampRequest tsq, Org.BouncyCastle.Tsp.TimeStampResponse tsr){
/* This compares the nonce and message imprint and whatnot in the TSTInfo.
* It throws an exception if they don't match. This doesn't validate the
* certs or signatures, though. We still have to do that in order to trust
* this data.
*/
tsr.Validate(tsq);
var tst = tsr.TimeStampToken;
var timestamp = tst.TimeStampInfo.GenTime;
var signers = tst.ToCmsSignedData().GetSignerInfos().GetSigners().Cast<Org.BouncyCastle.Cms.SignerInformation>();
var certs = tst.GetCertificates("Collection");
foreach(var signer in signers){
var signerCerts = certs.GetMatches(signer.SignerID).Cast<Org.BouncyCastle.X509.X509Certificate>().ToList();
if(signerCerts.Count != 1)
throw new Exception("Expected exactly one certificate for each signer in the timestamp");
if(!signerCerts[0].IsValid(timestamp)){
/* IsValid only checks whether the given time is within the certificate's
* validity period. It doesn't verify that it's a valid certificate or
* that it hasn't been revoked. It would probably be better to do that
* kind of thing, just like I'm doing for the signing certificate itself.
* What's more, I'm not sure it's a good idea to trust the timestamp given
* by the TSA to verify the validity of the TSA's certificate. If the
* TSA's certificate is compromised, then an unauthorized third party could
* generate a TimeStampResp with any timestamp they wanted. But this is a
* chicken-and-egg scenario that my brain is now too tired to keep thinking
* about.
*/
throw new Exception("The timestamp authority's certificate is expired or not yet valid.");
}
if(!signer.Verify(signerCerts[0])){ // might throw an exception, might not ... depends on what's wrong
/* I'm pretty sure that signer.Verify verifies the signature and that the
* signed attributes contains a hash of the TSTInfo. It also does some
* stuff that I didn't identify in my list above.
* Some verification errors cause it to throw an exception, some just
* cause it to return false. If it throws an exception, that's great,
* because that's what I'm counting on. If it returns false, let's
* throw an exception of our own.
*/
throw new Exception("Invalid signature");
}
}
}
static byte[] ReadFromBuild(string key){
// not shown
}
I am not sure to understand why you want to rebuild the data structure signed in the response. Actually if you want to extract the signed data from the time-stamp server response you can do this:
var tsr = GetTimestamp(hashToTimestamp, nonce, "http://some.rfc3161-compliant.server");
var tst = tsr.TimeStampToken;
var tsi = tst.TimeStampInfo;
var signature = // Get the signature
var certificate = // Get the signer certificate
var signedData = tsi.GetEncoded(); // Similar to tsi.TstInfo.GetEncoded();
VerifySignature(signedData, signature, certificate)
If you want to rebuild the data structure, you need to create a new Org.BouncyCastle.Asn1.Tsp.TstInfo instance (tsi.TstInfo is a Org.BouncyCastle.Asn1.Tsp.TstInfo object) with all elements contained in the response.
In RFC 3161 the signed data structure is defined as this ASN.1 sequence:
TSTInfo ::= SEQUENCE {
version INTEGER { v1(1) },
policy TSAPolicyId,
messageImprint MessageImprint,
-- MUST have the same value as the similar field in
-- TimeStampReq
serialNumber INTEGER,
-- Time-Stamping users MUST be ready to accommodate integers
-- up to 160 bits.
genTime GeneralizedTime,
accuracy Accuracy OPTIONAL,
ordering BOOLEAN DEFAULT FALSE,
nonce INTEGER OPTIONAL,
-- MUST be present if the similar field was present
-- in TimeStampReq. In that case it MUST have the same value.
tsa [0] GeneralName OPTIONAL,
extensions [1] IMPLICIT Extensions OPTIONAL }
Congratulations on getting that tricky protocol work done!
See also a Python client implementation at rfc3161ng 2.0.4.
Note that with the RFC 3161 TSP protocol, as discussed at Web Science and Digital Libraries Research Group: 2017-04-20: Trusted Timestamping of Mementos and other publications, you and your relying parties must trust that the Time-Stamping Authority (TSA) is operated properly and securely. It is of course very difficult, if not impossible, to really secure online servers like those run by most TSAs.
As also discussed in that paper, with comparisons to TSP, now that the world has a variety of public blockchains in which trust is distributed and (sometimes) carefully monitored, there are new trusted timestamping options (providing "proof of existence" for documents). For example see
OriginStamp - Trusted Timestamping with Bitcoin. The protocol is much much simpler, and they provide client code for a large variety of languages. While their online server could also be compromised, the client can check whether their hashes were properly embedded in the Bitcoin blockchain and thus bypass the need to trust the OriginStamp service itself.
One downside is that timestamps are only posted once a day, unless an extra payment is made. Bitcoin transactions have become rather expensive, so the service is looking at supporting other blockchains also to drive costs back down and make it cheaper to get more timely postings.
Update: check out Stellar and Keybase
For free, efficient, lightning-fast, widely-vetted timestamps, check out the Stellar blockchain protocol, and the STELLARAPI.IO service.