On Tuesday, 01/30/2007 at 05:07 CET, Ken Vance <[EMAIL PROTECTED]> wrote:
> Hi, 
> 
> We have five VM LPARS on our z990.  One of these LPARS, LPAR 1, has a 
CTC to a 
> z800.  The CTC is dedicated to the TCPIP running on the LPAR, and it 
connect to 
> the TCPIP running on the z800.  This is an interim solution until we get 
the 
> z800 OSA connectivity established.  I wanted to route all z990 IP 
traffic 
> destined for the z800 to LPAR 1, and then over to the z800.  The only 
> connection from all LPARS to LPAR 1 is a HiperSocket connection.  I 
tried to 
> establish a connection via TCPIP, but I see that there is no option to 
> designate the LPAR 1 hipersocket device as a router.  If I was using an 
OSA 
> card, I would define it as a PRIROUTER.  This would enable it to accept 
other 
> IP addresses and forward them on other links.  If I look at the 
hipersocket 
> link, I see that it is defined as a NonRouter. 
> 
> Is there a way of enabling the hipersocket on LPAR 1 for IP forwarding?

Linux and z/OS are able to proxy a HiperSocket onto the real LAN by using 
their respective "HiperSocket Accelerator" functions (I assume they still 
exist).  But you've got too many hops for that to work AND you're hopping 
across CECs.  LAN -> OSA on z990 LPAR x -> z990 LPAR 1 TCPIP via 
HiperSockets -> CTC -> z800 TCPIP -> whatever.

IMO, it'll be far easier and faster to get the z800's OSA connection up 
and running that it will be to assign new IP addresses in new subnets and 
then figure out how to bend the routes to your will and get those routes 
distributed in the network.  You won't be able to use the same IP address 
for z800 TCPIP in the above configuration as you would use if z800 TCPIP 
were directly connected to the LAN.  Well, I guess you could probably use 
a 3-meter routing prybar, but it would be a nightmare and by the time you 
got it to work, the OSAs will be ready and you'll be covered in 
lacerations and contusions.  Not to mention sweaty and dirty.

(What is the inhibitor to getting the z800's OSAs online?)

Alan Altmark
z/VM Development
IBM Endicott

Reply via email to