Re: VxWorks and Radius Authentication

[Ralph Lange <Ralph.Lange@bessy.de>]

For the conserver-server, connecting to the soft IOC under procServ is exactly the same as connecting to a VME IOC connected to a terminal server - it opens a telnet connection to a port on the remote host.

If procServ only provides access from localhost, there are two options:
1. The conserver-server connects opening a ssh connection to the softIOC host using a key that is configured to call "telnet localhost nnnn" to connect it to the softIOC. (Or explicitly calling "telnet localhost nnnn" on the other side.)
2. The softIOC host runs a conserver-server locally, which can simply run telnet to get to the softIOC.

conserver supports configurations with multiple conserver-servers.
Its client (called console) only needs the IP of one of the conserver-servers. When console connects, the conserver-server redirects it to the instance that actually hosts the connection the clients wants to get to.

So, on any client machine, typing "console <ioc-name>" connects you to that IOC's console, no matter which conserver-server hosts the connection, no matter if it is a soft IOC or a VME box.

The main advantages of putting conserver on top are:
- One way to connect to softIOCs and VME IOCs, no configuration on the client machine (other than the IP of one conserver-server)
- Authentication / Authorization: password-based auth, allows specifying user groups with different access rights on different consoles
- One way of logging all console output (with timestamps) to file or into the syslog protocol
- Replay buffer that can show the console activity before you connect
- Only one user at a time has write access, all others are read-only, but users with matching authorization may take over a console

Ralph


05.08.2010 15:26, Stephen Lewis wrote:

How does conserver fit it with procServ? ÂSeems like it is harder to use and very centralized. ÂI use procServ and have a simple shell script that looks up the telnet host/port in a table by IOC name.

On 5 Aug 2010, at 12:11 , Ralph Lange wrote:

http://conserver.com/