On your left you see a picture which is a general illustration of the departments network topology. You will also see that the map has been split into two major areas,
a green and a gray. All of the connections in the gray are, what I call,
Hackable network connections. I call these connections hackable because all
these machines connect to the CWRUnet and the CWRUnet is connected to the rest
of the internet. This means that anyone in the world with a connection to the
internet has the ability to try connecting to any machine on CWRUnet or listen
to CWRUnet. Yes it is possible to write programs that listen to the network
traffic generated between computers, and if information crossing the network is
not encrypted then hackers can steal information. Favorite information for
hackers are usernames and passwords. If information like that is sent across the
network unencrypted, the hacker can then log in as you and do untold amounts of
damage to your information. This is why I require that windows users use a
program called Secure CRT. Secure CRT is a program that uses ssh to allow you to
connect to servers like paralogy. The ssh program encrypts all information being
passed between computers to prevent hackers from getting and, or modifying your
data, or still worse, crashing the server itself.
In an effort to make the biocluster (P2-P16) more secure it has been placed on a PN (Private Network) as illustrated by the blue network connections. The PN has absolutely no knowledge of the CWRUnet, and the CWRUnet has absolutely no knowledge of the PN. This is in part necessary to prevent the two networks from crashing one another, but it also has the effect of securing the PN from hackers. Since the PN is not connected to the internet hackers connected to the internet cannot attack the PN. This makes it possible to support all kinds of network protocols between the machines of the biocluster with no need of concern about security. This is why an insecure means (like telnet) of connecting to other computers is permitted here. Remember, these machines can only talk to each other.
The question then remains, if the machines on the PN cannot connect to the network how can you access them and how does it access shared resources like the blast database. The trick is paralogy's. There are two network cards in paralogy, allowing it to see and connect to CWRUnet and the PN simultaneously. As far as P2-P16 are concerned, paralogy is on their network and no other. As far as CWRUnet is concerned paralogy is on its network and no other. This allows both networks to talk to paralogy.
SO, How do you get from you machine connected to CWRUnet to a machine connected to the PN. You have to start by connecting to the one machine that can see both networks, and that is paralogy. You have to use an ssh client like Secure CRT to do this. Once you are logged into paralogy you then have to use it to jump to any of the other machines in the biocluster. To do this, you can use telnet to connect to any of the machines on the secure network. You do this by typing telnet <hostname>(for example: telnet P9) at your command line prompt on paralogy. You will then be asked to login again, and then you will be connected to P9 via paralogy. You will therefore be constrain by all the limitations of being on the PN, so programs that want to connect to ncbi (like fasta_download) will not work on P9 because P9 is not connected to the internet. If you want to use programs like fasta_download, you can either login and logout as you need to download files and start jobs, or you can login to paralogy twice and use one session to telnet to another processor and the other to download any information you need.