SSH Key generation question
in [General Linux]
|
From: Maxwell Lol on 20 Jun 2010 16:49 Todd <todd(a)invalid.com> writes: > I would have to point out that the workstation is the > one who initiate the network packet as state=new, not the > number cruncher. Wrong.
From: J G Miller on 20 Jun 2010 17:00 On Sun, 20 Jun 2010 13:35:17 -0700, Todd wrote: > You do an excellent job of explaining it. Only because when I first encountered X11 I did not understand why the local machine was the X11 server until it was carefully explained to me. > Regardless of the DISPLAY variable The DISPLAY variable is only relevant to X11 client programs, and maybe the Esound Server or Pulse Audio server. > the workstation is responding to output from the cruncher > that it requested. Actually, it is software on the workstation, not the hardware its-self ;) > The workstation is not blindly listening to anything. I think it is safe to say that its network port is listening to anything according to the network protocol, and at the ethernet level is listening to anything going past. > If I were to send blind rendering from another computer, the > workstation would reject them as I had not established a network > connect with it first. If you have not authenticated yourself with the database server, it will also refuse to provide you with any data. > The workstation's X portion is only acting as a server in a micro sense. Not in a micro sense, but in the perspective of the X11 protocol. > In a macro sense, it is acting as a client: it is responding to > something it requested. In the usual "macro" sense, clients send requests and servers respond to them. Of course, the client usually waits around for the results to be sent back. In an X11 client program, you normally have a main loop, so the client is always waiting for events (mouse clicks, mouse movements, key strokes etc) to process and then send to the server which translates them to images or image movements on the screen. And one of the very first things that the "on screen" part of an X11 program that a client does is connect to the X11 server using the DISPLAY environmental variable or -display parameter with Display *XOpenDisplay(char *display_name); Surely you would agree that the display *server* has to be there, just waiting for an authenticated client to connect? What it all comes down to is your frame of reference. The workstation is the display server in the context of X11, whereas the data files for processing could be on a file server on machine A, and the database server where they get stored is on machine B, and the processing result is sent to a print server on machine C. Rather than thinking of the hardware as the server, you should be thinking of what services the machine offers, since these are determined primarily by the software that is running, not the actual box.
From: J G Miller on 20 Jun 2010 17:04 On Sun, 20 Jun 2010 13:06:41 -0700, Todd wrote: > I am thinking that sending him his private key "out-of-band" (through > the postal service on a CD in a password scrambled zip file) would work > around the situation. It will take less time for the business client to fire up puttgen than it will to get the CD out of the case, mount it, find the contents in the file explorer, click on it, open the password, unzip the file and then worry about moving the contents to an appropriate directory. Furthermore, does the business client really want you to have the security of his account compromised by you holding a copy of his private key?
From: David W. Hodgins on 20 Jun 2010 17:04 On Sun, 20 Jun 2010 15:52:56 -0400, Todd <todd(a)invalid.com> wrote: > Big picture: the workstation is listening for rendering because > of something it requested from another computer. Therefore > it is client. Here's an example that may help make it clearer. I use ssh to establish a connection to a remote computer. On that computer, I start gkrellm. The client, gkrellm requests a connection to the x11 server running on my machine, and uses that connnection to send the dialog to the server, and get keyboard and mouse input, from the server. The server controls the hardware, and only responds to client requests. The client can be on the same computer, or a different one. Regards, Dave Hodgins -- Change nomail.afraid.org to ody.ca to reply by email. (nomail.afraid.org has been set up specifically for use in usenet. Feel free to use it yourself.)
From: Todd on 20 Jun 2010 17:52
On 06/20/2010 01:16 PM, J G Miller wrote: > On Sun, 20 Jun 2010 12:52:56 -0700, Todd complained: > > The DISPLAY environmental variable is the key to understanding which > X11 server it is that you are using. > > If it is just :0.0, it is using the local X11 server. If it is > set to this_workstation:0.0 then it using the local X11 server. > If it is set to some_other_machine:0.0 then output of the X11 > clients are going to the X11 server on that some_other_machine. The question arises are to who request the display variable be set that way? Not the cruncher. The cruncher is only responding to what was requested of it. In the micro sense of things, the workstation's X portion could be construed as being a server as it is responding to output from the cruncher. My beef: the functional result is what they should call client and server. By the way, vncserver uses the same DISPLAY variable and could have played cute with that along the same lines as X11. Thank goodness they did not, or no one would be able to tell what was going on. My point: name your client and server based on the functional outcome, not one barrier process. -T |