Anybody know how https *really* works? I didn't think so

Do you have a question? Post it now! No Registration Necessary.  Now with pictures!

Threaded View
So my book on https and Windows Communication Foundation technology
says that if any computer between your SSL certificate secured
computer and the client machine reading this certificate does not
support SSL, then the entire https link is not secure and your data
can be compromised.  That makes no sense to me, because I thought the
entire data stream is encrypted, but that's what it says.  And I've
even seen this on the net.

So why do people blindly trust SSL and HTTPS as if it's unbreakable?
Is it because most traffic goes through at most two or three hops, and
likely these hops are up-to-date and support SSL?

Even if so, you're taking a risk that somewhere between somebody will
fail to support SSL and your message will be unencrypted.

Bet most if not all of you reading this thread did not know this.  So
called experts, right.

RL

Re: Anybody know how https *really* works? I didn't think so


Do you know how Usenet *really* works?  I didn't think so.


--
Dave
Multi-AV Scanning Tool - http://www.pctipp.ch/downloads/dl/35905.asp



Re: Anybody know how https *really* works? I didn't think so

+1

--
Peter
Please Reply to Newsgroup for the benefit of others
Requests for assistance by email can not and will not be acknowledged.
This posting is provided "AS IS" with no warranties, and confers no rights.
http://www.microsoft.com/protect


Quoted text here. Click to load it


Re: Anybody know how https *really* works? I didn't think so

Quoted text here. Click to load it

Encryption is only as secure as the key management system is.

Quoted text here. Click to load it

Because they don't understand security as it pertains to encryption (or
vice versa).

Quoted text here. Click to load it

???

Quoted text here. Click to load it

It just unencrypts, like that?

...I don't think so.

Quoted text here. Click to load it

I don't think that you really fully understand what you are talking
about, so it seems ironic when you lamely attempt to insult and troll
those you somehow believe to be *experts* in so many disparate
crossposted groups.



Re: Anybody know how https *really* works? I didn't think so

wrote:
Quoted text here. Click to load it

Nope, Shiite->4Brains, that' s NOT what we are talking about.  Try
again.  We are talking about HTTPS, not key management.  Yes, it's
true that key management is only as secure as the lock on your door to
the secondary storage holding said keys, but again, that's not at
issue here.

Quoted text here. Click to load it

Right. ???.  That's your value add to this debate: ???.  That should
be your middle name: ???  The Reflex.

Quoted text here. Click to load it

Yes, just like that.  What you fail to understand (among your many
other failures) is the difference between message level security and
transport level security.  HTTPS is the latter not the former.  Here's
a reference for you to 'bone up' on, bonehead: (http://
msdn.microsoft.com/en-us/library/ms733137%28VS.90%29.aspxEnd-to-end
security. A secure transport, such as Secure Sockets Layer (SSL) works
only when the communication is point-to-point. If the message is
routed to one or more SOAP intermediaries before reaching the ultimate
receiver, the message itself is not protected once an intermediary
reads it from the wire. Additionally, the client authentication
information is available only to the first intermediary and must be
transmitted to the ultimate received in out-of-band fashion, if
necessary. This applies even if the entire route uses SSL security
between individual hops. Because message security works directly with
the message and secures the XML in it, the security stays with the
message regardless of how many intermediaries are involved with the
message before it reaches the ultimate receiver. This enables true end-
to-end security scenario.=94)

Quoted text here. Click to load it

You *STILL* don't think so, even after reading the above?  Man youz
stupid.

Quoted text here. Click to load it

NOT.  I hope you lerned something from this thread, dopehead.

Anybody else?  C'mon down!  Insults are free of charge.

RL

Re: Anybody know how https *really* works? I didn't think so


Quoted text here. Click to load it

The only thing left to debate--and I doubt the small minds in this
group has the capacity to address this issue (no thanks in advance)--
is how often "SOAP intermediaries" are present in a 'typical' message
route.  I would bet that for most 'routine' messages such as home user
to bank server, there would be no intermediaries, and the ISP server
is just "pass through" and would not require SOAP (I would imagine).
But this is a question for a real expert, not the dunces that hang
around the virtual water cooler that passes for Usenet these days.

RL

Re: Anybody know how https *really* works? I didn't think so

wrote:
Quoted text here. Click to load it

Nope, Shiite->4Brains, that' s NOT what we are talking about.

***
So it just magically unencrypts itself?
***

Try again.  We are talking about HTTPS, not key management.  Yes, it's
true that key management is only as secure as the lock on your door to
the secondary storage holding said keys, but again, that's not at
issue here.

***
So it just magically unencrypts itself?
***

Quoted text here. Click to load it

Right. ???.  That's your value add to this debate: ???.  That should
be your middle name: ???  The Reflex.

***
That was an indication that I didn't understand what you were talking
about, but I see now what that was so.
***

Quoted text here. Click to load it

Yes, just like that.  What you fail to understand (among your many
other failures) is the difference between message level security and
transport level security.  HTTPS is the latter not the former.

***
Why the animosity? Can you explain how something can encrypt at one end,
and decrypt at the other, without some kind of key being involved?
***

[...]

Quoted text here. Click to load it

You *STILL* don't think so, even after reading the above?  Man youz
stupid.

***
Not really, it's just that you not only fail to make sense, you fail to
understand the subject enough to explain to me what you actually meant.

...and you're still acting like an asshole toward me for no good reason
(aside from the obvious trolling that is).
***

Quoted text here. Click to load it

NOT.  I hope you lerned something from this thread, dopehead.

***
Yep, I learned that you are a stupid troll.

Bye-bye
***



Re: Anybody know how https *really* works? I didn't think so

wrote:

Quoted text here. Click to load it

Nope.  You fail to understand how transport security works.  The one
passage that was not flamebait you clipped (and I reproduce it again,
below).

So, where you trolling then?  You clearly have no interest in lerning
anything from this thread, trollfeeder.

And I don't know why SOAP intermediaries break https.  That really was
my question to the group.

Bye.

RL

(http://
msdn.microsoft.com/en-us/library/ms733137%28VS.90%29.aspxEnd-to-end
security. A secure transport, such as Secure Sockets Layer (SSL)
works
only when the communication is point-to-point. If the message is
routed to one or more SOAP intermediaries before reaching the
ultimate
receiver, the message itself is not protected once an intermediary
reads it from the wire. Additionally, the client authentication
information is available only to the first intermediary and must be
transmitted to the ultimate received in out-of-band fashion, if
necessary. This applies even if the entire route uses SSL security
between individual hops. Because message security works directly with
the message and secures the XML in it, the security stays with the
message regardless of how many intermediaries are involved with the
message before it reaches the ultimate receiver. This enables true
end-
to-end security scenario.=94)

Re: Anybody know how https *really* works? I didn't think so

RayLopez99 wrote:
Quoted text here. Click to load it

Perhaps Microsoft didn't explain it well enough for you.

Design 1. C (client) -- internet -- Z (destination server)

If your client uses HTTPS and the URL of Z, then your message is safe.

Design 2. C -- internet -- S (intermediary) -- internet -- Z

If your client uses the URL of S, then S uses the URL of Z (even if
they're both using HTTPS) then your message may be read/altered by S.

What was not said is that in design 2, S should really be considered C's
destination, and Z is S's destination - and that protocol encryption
(HTTPS) only protects your message on its path through the internet.

If you don't want S to be able to read/alter your message then encrypt
the message so that only Z can read it - or use design 1 and HTTPS.

Re: Anybody know how https *really* works? I didn't think so

wrote:
Quoted text here. Click to load it

Neither did you Jason, but I appreciate the attempt.  Please read on
however.

Quoted text here. Click to load it

What are you saying here?  Are you saying that S has the ability to
"alter" your message, by say garbling it?  To make it unreadable?
(That is, change every letter Y to X and Z character to A, etc?  That
would be simple vandalism, and not really a security 'breach' in my
mind.  Or are you saying that S has a private key to the HTTPS and can
unencrypt your encrypted message?  That was what I thought originally--
and it's still not clear how S can get a private key--only "Z" has
such a key (that's my understanding).   I thought HTTPS uses some form
of asymmetric public key (I trust you know what this is), and that the
only holder of the private key is Z.  But if HTTPS uses a symmetric
key, then I can see how S can indeed decrypt the message from C and
read it.    Please explain.  That's the last time I use "please" in
this thread BTW.

Quoted text here. Click to load it

Again, that is how I thought things work:  using a asymmetric key,
that's exactly how things should work:  every two points in a chain of
transmission is as strong as the next two points--there are no weak
links.

Quoted text here. Click to load it

RL

Re: Anybody know how https *really* works? I didn't think so

RayLopez99 wrote:
Quoted text here. Click to load it

As I said, in design 2, C is really communicating with S. If you wrote
S, then all is good - you can probably trust it. However, the service S
receives the original (decrypted) message - because its URL was used by C.

S's job is to communicate with Z. If it uses HTTPS to do that, then the
original message is again (protocol) encrypted and sent on its way to Z.

All anyone is saying is that because C was "talking" to S (using HTTPS),
and not directly to Z, then S gets to see/read the original message
(assuming no message encryption before being posted by C).

Possible reasons for having an intermediary in the process include:
- load balancing (there could really be several Z's)
- S's job may be to translate/transpose the original message into a form
that Z can understand
- extra security: the firewall between C and S may be configured
differently to the one between S and Z

RayLopez99 had a hard date last night. That's why he can't sit down.

Don't worry, a package of suppositories are on their way.

Re: Anybody know how https *really* works? I didn't think so

wrote:
Quoted text here. Click to load it

Are you sure of this?  You use the term "load balancing" which is a
term of art database engineers use, so I take it you have some
knowledge of this area, but frankly your suggestion that what people
are saying about HTTPS and intermediaries is that the intermediary can
decrypt an HTTPS stream and read the unencrypted message is a bit daft
to me.

Here's why:  if S, the intermediary, has the ability to decrypt the
SSL certificate, it means it also has a private key (like Z does).
That means at some point Z, the final destination endpoint, gave S the
private key to decrypt message streams--otherwise how could S do it?
Or perhaps C, the client, did.  In either case somebody at either
endpoint had to 'trust' S.  If that trust was misplaced, and it turns
out S is a crook, well that's life--but somebody made the conscious
decision to trust S.

Does this make sense?  So 'human error' in trusting an intermediary
like S is what Microsoft is talking about?  That should be made more
clear if so.

RL

RayLopez99 had a hard date last night. That's why he can't sit down.

He said load.

Re: Anybody know how https *really* works? I didn't think so

RayLopez99 wrote:
Quoted text here. Click to load it
<snip>
Quoted text here. Click to load it

No. Private keys are never shared with anyone.

Further reading:
http://www.ourshop.com/resources/ssl.html

Re: Anybody know how https *really* works? I didn't think so

Quoted text here. Click to load it

It may take a while, but eventually he will understand that the data is
"covered" only in transit just as with TPM supported disk encryption it
is only covered at rest.



Re: Anybody know how https *really* works? I didn't think so


Quoted text here. Click to load it

Explain please in simple words.  You are saying encryption of https
data is only done in transit?  Why/how?  To be more explicit:  C
(client) --> S (intermediary) --> Z (endpoint/host server):  you are
saying encryption is only at the "-->" in between the points, C, S and
Z?  That is, at the servers C,S,Z you can read the data packets in
plaintext, but when they transmit the data stream they become
encrypted ciphertext?

If so, this does make Jason Keats explanation work, but I would be
very surprised if this is so.  Why would anybody design an algorithm
that decrypts as soon as it reaches a server?  If routing is the
reason, you can (I would suppose) just keep the headers decrypted and
route the message using the headers, which is conventional.

Please confirm.  That's the second time I've used please in this post
and twice too many.

RL

Re: Anybody know how https *really* works? I didn't think so

On Mon, 1 Nov 2010 10:08:49 -0700 (PDT), RayLopez99 wrote:

Quoted text here. Click to load it

See what I mean? Pants on fire there, assclown.
--
http://tr.im/1fa3

Re: Anybody know how https *really* works? I didn't think so


Quoted text here. Click to load it

Explain please in simple words.  You are saying encryption of https
data is only done in transit?  Why/how?  To be more explicit:  C
(client) --> S (intermediary) --> Z (endpoint/host server):  you are
saying encryption is only at the "-->" in between the points, C, S and
Z?  That is, at the servers C,S,Z you can read the data packets in
plaintext, but when they transmit the data stream they become
encrypted ciphertext?

***
The client and server have a trust relationship, they agree on a keyset
and the data is encrypted by the client, sent out on the wire, received
by the server, and decrypted by the key. If it needs to do it again (not
the final destination), another trust relationship is set up btween the
server and the next step and the process may be repeated (new
relationships, new keys, no transitive trust). The idea was to have the
data covered so that sniffing or otherwise capturing packets enroute
would not have value to the interloper (the encryption security is as
good as the key management - if the interloper (you don't have a trust
relationship with) doesn't have the key, he cannot convert the
ciphertext to plaintext.

You are *trusting* all of the SSL capable servers with your data unless
you "cover" your data for the entire source to destination trip.
***

If so, this does make Jason Keats explanation work, but I would be
very surprised if this is so.  Why would anybody design an algorithm
that decrypts as soon as it reaches a server?

***
Those whom are only concerned with the security (integrity) of the
point-to-point communications.

Think of the old can-to-can voice communication, and someone
eavesdropping with a tap on the string. You still want to be able to
speak into one end and hear from the other, but you want the
eavesdropper to get nothing intelligible. If your can-to-can connection
is only one part of a relay communication network, you wouldn't be
concerned about the relay personnel actually having access to that
information uncovered. If you were concerned about that, then end-to-end
coverage is what you really want.
***



Re: Anybody know how https *really* works? I didn't think so

Quoted text here. Click to load it

Thanks.  OK, but who sets up the 'new keys', 'new relationships'?  It
has to be done by somebody using new SSL certificates, not the
original one used by C?

Here is where I'm going:  C wants to send to Z.  It types in Z's URL
using an SSL certificate.  By the nature of the internet, inbetween C
and Z there are always other servers that act as relays, call one of
them 'S'.  Is it possible for them to use 'new keys', 'new
relationships' that would compromise the message from C to Z?  Unknown
to C?  I don't think so--or I can't see how.  That is, C would have to
send a message to 'S', not 'Z'.  But at that point, we are arguing
over semantics--'S' now *is* 'Z'!  Of course if S wants to send to Z,
that will mean S can read the message--but that's just plain stupid
semantics.

Please explain if it's possible for C to send to Z, then, if there's
an intermediate SOAP server S, unknown to C (maybe not unknown, but
perhaps unsuspected for a security risk) whether S can decrypt the
message if C has typed in Z's URL.  Unless--and maybe this is
possible--C types in Z's URL but the program--behind the scenes--
changes Z's URL to S!  Why would it do that--can it do that?

RL

Site Timeline