Thanks for the feedback.
If I might add a little more...
We're presented with the challenge of equitable billing to our clients that
have applications running on WebSphere. Ideally, they want to see a flat
rate presented to them
on a periodic basis (monthly, quarterly, etc) for planning/budgeting
purposes. If all applications were the same and consumed the same amount
of resources, it wouldn't be so difficult. We could identify our costs and
divide them equally, assuming some number of applications running on a
single virtual machine. That's far from reality for us and very likely
others as well. So we try to add the usage factor in the mix to be fair to
everyone. Current thinking is that we could establish some base rate or
'entry fee' to cover some amount of fixed costs regardless of the number of
apps and then present a periodic maintenance/usage fee based on cpu
utilization or hits, etc..
Based on the work I've done so far with WebSphere, there appear to be a
couple of different ways to configure the 'servers'. One option is to
install all applications under the 'default server' and the other is to
create multiple virtual servers on the same virtual machine. The latter
seems like it may be of greater benefit from an administration and
isolation perspective. I don't currently know how we would isolate cpu
usage for either of them given that all apps would be running on the same
virtual machine even though they are structured differently. Creating
individual webservers/virtual machines on our LinuxIFL doesn't seem to be
an option based on resource constraints. (Real storage - 2GB limit).
WebSphere uses a significant amount of resources on a per virtual machine
basis. We are exploring means of stretching the available resources in
order to facilitate additional virtual machines with WebSphere but it
doesn't look like we'll gain a great deal.
We have alot of tradition/history in the usage based billing and think that
it has a place here also. The challenge is how to come up with a method of
incorporating it into a scheme
and present something that our clients can use for budgeting purposes, yet
be fair to everyone. The alternative for us may be to break away from
history which also presents new
challenges.
|---------+---------------------------->
| | Rich Smrcina |
| | <[EMAIL PROTECTED]|
| | om> |
| | Sent by: Linux on|
| | 390 Port |
| | <[EMAIL PROTECTED]|
| | IST.EDU> |
| | |
| | |
| | 09/11/2003 08:38 |
| | AM |
| | Please respond to|
| | Linux on 390 Port|
| | |
|---------+---------------------------->
>---------------------------------------------------------------------------------------------------------------|
|
|
| To: [EMAIL PROTECTED]
|
| cc:
|
| Subject: Re: Web Server Chargeback/Cost Recovery Methods
|
>---------------------------------------------------------------------------------------------------------------|
Traditionally this is done through machine utilization information that
is logged by VM's accounting facility. Some value is assigned to a unit
of work (a CPU second) and machines that are to be charged back have
their accounting information tracked and billed to the user/customer.
The accounting facility is a standard deliverable with VM. I don't
believe that there are any native tools available to report in it, but
there are packages and products available to do this. Of course, you
could roll your own.
On Thu, 2003-09-11 at 08:36, Jerry Prange wrote:
> I'm interested in ideas on how to structure a billing/chargeback system
for
> webservers running on Linux390.
>
> Thank you.
--
Rich Smrcina
Sr. Systems Engineer
Sytek Services - A Division of DSG
Milwaukee, WI
rsmrcina at wi.rr.com
rsmrcina at dsgroup.com
Catch the WAVV! Stay for requirements and the free-for-all.
Update your zSeries skills in 4 days for a very reasonable price.
WAVV 2004 in Chattanooga, TN
April 30-May 4, 2004
For details see http://www.wavv.org
