DHCP performance testing

[Frank Bulk]

Puter:

My apologies for not answering your question more directly.

I checked out two vendors in this field, Nominum, BlueCat networks, and
Infoblox, and saw mention of leases per second
(http://www.nominum.com/getFile.php?file=nominum_ds_dcs.pdf &
http://www.bluecatnetworks.com/products/adonis-dns-dhcp-appliances/dns-dhcp-
management/techspec/ & https://infoblox.com/products/pdf/ds_hardware.pdf).
That said, it's not been issue of much discussion here.  There have been
discussions around how well ISC's DHCP responds in the event of a cable node
failure or building power outage, but generally there are other gating
factors for service restoration than DHCP, such as TFTP performance.

A worst case scenario is complete power outage with simultaneous power on.
It's unlikely that your 60,000 nodes would experience that, unless they
really are located on one campus.  I haven't seen one, but I hope that some
of the DHCP appliance vendors have done some research that calculates the
bell curve for lease requests with x nodes, and how quickly their request is
satisfied if the appliance can perform y leases per second.

Regards,

Frank

-----Original Message-----
From: dhcp-users-bounce at isc.org [mailto:dhcp-users-bounce at isc.org] On Behalf
Of Puter ami
Sent: Friday, July 13, 2007 6:34 PM
To: dhcp-users at isc.org
Subject: RE: DHCP performance testing

>From: "Frank Bulk - iNAME" <frnkblk at iname.com>

>As has been discussed on this forum many times before (just read the
>archives), performance is not often an issue for ISC's DHCP.

Thank you for the reply, Mr. Bulk.

Perhaps I did not make myself clear. I am not asking if DHCP is often an
issue. I have perused the archives. I am asking how evaluators setup a
common set of definitions and criteria for the setting of performance
specifications for a DHCP server.

For instance, with DNS, the tool and performance specifications commonly
used in the industry is "queryperf" and "queries per second (qps)".

My question is what tool and what performance specifications are commonly
used with DHCP? What is a good number to expect in an Enterprise-class
network?

Perhaps if I can be so bold to say that a "lease per second (lps)" is the
DORA process, what is a good "lps" value for a network with ~60K nodes in a
worst case scenario?

>The limiting
>factors tend to be disk write time (updating the lease file) and reading in
>configs (for very large operations).  This can be somewhat mitigated by
>using a flash drive rather than a traditional magnetic hard disk drive.

I do understand general concepts and how to tune/uipgrade a DHCP server. I
am needing numbers (and definitions of those numbers) to give to prospective
vendors so they may recommend the hardware and software, then apply those
numbers in a laboratory environment and test the claim.

>Regards,
>
>Frank

Ralph Bischof

>-----Original Message-----
>From: dhcp-users-bounce at isc.org [mailto:dhcp-users-bounce at isc.org] On
>Behalf
>Of Puter ami
>Sent: Thursday, July 12, 2007 4:27 PM
>To: dhcp-users at isc.org
>Subject: DHCP performance testing
>
>Hello,
>
>I am tasked with determining needs for various components for a very large
>organization segmented to large centers. I have been trying to research how
>to size and test for the DHCP implementation. None of the information I
>have
>been finding on the web gives me meaningful numbers.
>
>I am asking for help in an industry-standard way of talking/testing DHCP
>performance. Or something that is industry-accepted. In other words, I need
>a definition and a level of performance to provide to companies so that
>they
>may answer with the type/size/number of hardware pieces (as well as
>software) to accommodate my customer's needs. This is for an
>Enterprise-class network.
>
>I have heard "operations-per-second", "transactions-per-second", etc. How
>do
>people define and measure performance? I am most interested in what server
>performance is nominal for a large network (let's say ~60K nodes) for
>average and peak (everybody turns on at 8 in the morning).
>
>Any pointers/help is much appreciated!
>
>Ralph Bischof

_________________________________________________________________
http://liveearth.msn.com