Linking Wired and Wireless network...

Michael D. Setzer II mikes at kuentos.guam.net
Sun Mar 13 23:50:07 UTC 2005


On 13 Mar 2005 at 13:26, Les Mikesell wrote:

From:           	Les Mikesell <lesmikesell at gmail.com>
To:             	For users of Fedora Core releases <fedora-list at redhat.com>
Date sent:      	Sun, 13 Mar 2005 13:26:16 -0600
Subject:        	Re: Linking Wired and Wireless network...
Send reply to:  	For users of Fedora Core releases <fedora-list at redhat.com>
	<mailto:fedora-list-request at redhat.com?subject=unsubscribe>
	<mailto:fedora-list-request at redhat.com?subject=subscribe>

> On Sun, 2005-03-13 at 00:11, Michael D. Setzer II wrote:
> 
> > I did a test after purchasing my own wireless card. In downloading 
> > the same file from the same remote site with a wired and wireless 
> > machine. The wireless machine got 40K per second, while the wired 
> > machine got 8K per second. 
> 
> This is probably a temporary situation and will change as more
> users start to share the wireless link.
> 

Not likely, the 3 wireless DSL lines are totally separate from wired 
lan, so only internet traffic without having to go out to the ISP and 
back in thru the T-1 line. Classrooms are only hard-wired, so 
classrooms could only change by adding wireless cards, and would 
then lose access to local servers and printing, 

> > The campus is setup by MIS with 4 Class C networks running on a 
> > single physical network (no subnets, no routing.) Only router is the 
> > one that connects to the T-1 line, and it has a 10MB ethernet 
> > interface. Three of the 4 networks are in the dhcpd pool, so I end up 
> > with classroom have a mix of all 3 networks. Without added routing 
> > entries, traffic would have to go thru the 10MB rounter instead of 
> > directly. 
> 
> If your 4 class C's are adjacent and aligned on the right
> bit-boundary they could be supernetted with a netmask
> of 255.255.252.0.  That won't help with internet access,
> but machine<->machine communication would not need to
> bounce though the router.  I'd be surprised if that isn't
> the case already. 
> 

11001010.10000000.01000111.00000000
11001010.10000000.01001000.00000000
11001010.10000000.01001001.00000000
11001010.10000000.01001111.00000000

Those are the 4 IP blocks. 71, 72, 73, and 79. 
It would take a mask of 240 to get them together I believe. 


> > So, working with MIS isn't an option at this point. I was thinking of 
> > adding a wireless card to my fedora machine which also runs a 
> > squid proxy server used by my lab. 
> > 
> > Is there something that would help this, or am I wasting my time..
> 
> If most of the lab communication is workstation<->server you
> could dual-nic the server with a NAT configuration and use a
> local switch for the lab workstations.  Adding a wireless
> link would work, but in this situation it sounds it would
> be even better if someone could set up a larger caching
> proxy that your squid and others could configure as a parent
> to share the cache.
> 

Basically, the network is more Internet and Printer sharing. Students 
have no access to the servers except for printing (all students log in 
user "user" as the id with no password. 

I have a squid proxy server running on my Fedora machine, and 
was getting a 40% cache-hit ratio, But MIS didn't like it, since 
monitoring traffic at the router only showed my machine. Taking the 
extra step of checking traffic to my machine was to much work. 

That Fedora machine also has 9 ethernet ports, since the setup had 
been with each classroom having its own segment before MIS took 
over. Now, I get to use the machine to demonstrate 1 day a 
semester for 1 class. It has dhcp, nat, iptables, squid proxy, but right 
now only 1 machine connects to the ports in my office. 

Thanks again.

> -- 
>   Les Mikesell
>    les at futuresource.com
> 
> 
> -- 
> fedora-list mailing list
> fedora-list at redhat.com
> To unsubscribe: http://www.redhat.com/mailman/listinfo/fedora-list
> 


+----------------------------------------------------------+
  Michael D. Setzer II -  Computer Science Instructor      
  Guam Community College  Computer Center                  
  mailto:mikes at kuentos.guam.net                            
  mailto:msetzerii at gmail.com
  http://www.guam.net/home/mikes
  Guam - Where America's Day Begins                        
+----------------------------------------------------------+

http://setiathome.berkeley.edu
Number of Seti Units Returned:  15,861
Processing time:  30 years, 128 days,  7 hours, 27 minutes
(Total Hours: 265,879)





More information about the fedora-list mailing list