× The internal search function is temporarily non-functional. The current search engine is no longer viable and we are researching alternatives.
As a stop gap measure, we are using Google's custom search engine service.
If you know of an easy to use, open source, search engine ... please contact support@midrange.com.



Yeah, in the i world it's almost always about redundancy. I agree with the statements in this thread about Ethernet line utilizations approaching zero, especially in the GbE space.

The magic is *VIRTUAL IP. The concept is that at the Ethernet level communications isn't about IP it's about the MAC address of the Ethernet line. Each of your physical lines (1, 2, 3, more?) has a unique MAC address. Each of them also gets a fixed IP that is otherwise not used( for example 10.1.1.1, 10.1.1.2, 10.1.1.3. etc) . The IP address assigned to the *VIRTUALIP has no home initially, assume it's 10.1.1.100. When someone out there says 'Who the heck is 10.1.1.100?' (An ARP request) IBM i picks one of the boys form 10.1.1.1, 2, and 3 to 'proxy' for the 10.1.1.100 address (say 10.1.1.1). It then sends out it's MAC address in the response. All connections at that point forward to 10.1.1.100 go to the same MAC address as 10.1.1.1 and IBM i handles the inbound packets.

This is awesome until the Ethernet line for 10.1.1.1 gets pulled out. Now normally nobody would know what to do to find 10.1.1.100 as they would continue sending packets to the MAC address of 10.1.1.1 but it's down. Until their ARP cache ages out they would be stuck. However IBM i says, 'Hmmm, 10.1.1.1 just went boom, and I've hung 10.1.1.100 there. I better pick a new home' So perhaps it picks 10.1.1.2 and then sends out a 'gratuitous ARP' (that is, one nobody asked for) that says, '10.1.1.100 is at MAC Address(of 10.1.1.2)'

All the machines in the network that have an entry for 10.1.1.100 update it and immediately begin using the updated entry.

All of this happens so fast that nobody even notices, connections remain up, even 5250 sessions continue without delay.

Hope that helps!!

- Larry "DrFranken" Bolhuis

On 8/2/2011 1:50 PM, rob@xxxxxxxxx wrote:
I think he is talking about binding 1 IP to multiple NIC's. The purpose
being twofold:
- Redundancy
- Load balancing

Well, it would be a rare day when I would come close to tickling the
capability of a 1GB card. According to what I use the only card that ever
sees a significant load is a 100MB card on one lpar. All the rest are GB.

Not sure about the redundancy. How would one get redundancy? Ok, I can
see the card, and/or the cable from the card to the switch. However I
can't see multiple switches and stuff. How could one IP address be on
multiple switches? This new huge switch you could have ports on a couple
of different cards and perhaps they are hot swappable (a coworkers job).


Rob Berendt

As an Amazon Associate we earn from qualifying purchases.

This thread ...

Replies:

Follow On AppleNews
Return to Archive home page | Return to MIDRANGE.COM home page

This mailing list archive is Copyright 1997-2024 by midrange.com and David Gibbs as a compilation work. Use of the archive is restricted to research of a business or technical nature. Any other uses are prohibited. Full details are available on our policy page. If you have questions about this, please contact [javascript protected email address].

Operating expenses for this site are earned using the Amazon Associate program and Google Adsense.