Capacity Management

[UKVIKING]

As I usually only do Service Support, the requirement to implement Capacity Management (Network) for the customer was an interesting affair

I learned that Change and Capacity both need a good configruation and a well defined project management process

as well as good links to sales (vendor side) for getting additional capapcity in a timely manner

scope creep of course is a given
_________________
John Hardesty
ITSM Manager's Certificate (Red Badge)

Change Management is POWER & CONTROL. /....evil laughter

[UrgentJensen]

I hear there's a lot of creeps in Capacity.

UJ
_________________
Did I just say that out loud?

(Beige badge)

[asrilrm]

Capacity Management, hmm ...
I've got some experience in capacity measurement and planning, especially in the mainframe environment.
It was a very interesting activity to work on capacity measurement.
I remember we used CA-JARS to measure CPU usage, DASD IO queue, memory utilization, etc.
But that was the technical side. And one thing, you need to have monitoring tools to provide accurate measurements to feed to capacity planning.

[BorisBear]

I find that the bit that IT Departments/Suppliers are generally not good at is Business Capacity Management - especially for something like network.

Establishing Business Metrics of Importance forsomething like a network can be a bit of a mare (depends how complex your infrastructure is and what the users actually use).

[UKVIKING]

the the customer that I was doin change and capacity

I established (re-establihed actually) a Capacity Board on a weekly basis

as we used ehealth for reporting on utilization, latency etc, we used that tool to produce reports

As the network consists of dedicated point to point circuit - slowly being replaced by MPLS, I reported on 2 distinct aspects

1 - the B/W utilization for all the links - 40% was the cutoff as each site had 2 diverse links. The customer wanted to make sure that if one fails the traffic can be routed over the other w/o hitting 100%

2- the B/W for the MPLS links - which the design / technoology allows exceeding assigned b/w. Dont want to get into tech discussion - but MPLS allows exceeding paid for b/w

The action from the mtgs primarily were -

1 - get the network upgrade project team to give better projections & updates on link upgrades
2 - make recommendation to the network project teams to have upgraded or upgradign circuits to have b/w increase due to increase load way beyond the benchmark - note this supposedly would generate more income for the carrier... but...

The customer - started to understand capacity management (more utilization and projections...) and started to demand more and more info which could not be given due to the poor architectual design of the network

shared vlans, bad naming convention - poor layered architecture


but it was fun
_________________
John Hardesty
ITSM Manager's Certificate (Red Badge)

Change Management is POWER & CONTROL. /....evil laughter

[BorisBear]

[UKVIKING]

BorisBear,

The customer - a financial services company - had a vast global network of point to point circuits going from its major regions hub (EMEA, Americas, Asia Pacific). They had major trunks between the regions and modest size ones in region (depending on the business)

The carrier managed the network and worked with the customer to chaneg all the p2p to MPLS. The initial baseline was done 12 months before the project went in.

geek speak
MPLS consists of a physical link (access speed) and a logicial link (port speed). The physical links takes 3- 6 months to plan / install / connect. The physical link would go from the closest MPLS network point of presence to the customer site. Logical link takes no time at all.

Physical links are 2 mb, 34mb, 45mb, 155, 1Gb.
Logical can be anything up to physical speed

Hopefully you see the planning issue (regarding capacity)

The project to replace the p2p also coincide with a save money project

Therefore when the baseline circuit set was for a site to be X / Y and since then the utilization is creeping up due to increased business (opening of new services, etc (unbeknowst to the network people)), when I did the 'tactical capacity' I had to point out the project baseline issues with the actual util

This of course caused issues ont he project side and the customer

More ltr
_________________
John Hardesty
ITSM Manager's Certificate (Red Badge)

Change Management is POWER & CONTROL. /....evil laughter

[UKVIKING]

Boris Bear

Basically the answer regarding the BMI is no

I and the customer rep were designing the process as we went

I started with the obvious - network utilziation

now for a correction - the network was not poorly designed per se... the documentation evidence like most companies was sparse and kept under local knowledge control. key individuals had the knowledge - not documented

the good news was that the initiative by myself ...started to filter into the other environments as well
_________________
John Hardesty
ITSM Manager's Certificate (Red Badge)

Change Management is POWER & CONTROL. /....evil laughter

[MSP]

Viking,

First of all, thanks for the diversion to the Service Delivery / Capacity Management area. I think your experience very much aligns to the challenges I have had to deal with in the past.

I will give some insight into one of my most recent challenges, where I was partly responsible for almost 1000 user buildings dotted around the UK for one of the country's biggest service suppliers. The crux of our problem was:

a) We never really had a grasp on business capacity management. i.e. What is the business doing, when and how, that might impact our network.

b) our network utilisation was growing so fast, we couldn't deploy new infrastructure in time due to lack of testing and sluggish delivery dates.

As a result of this typical scenario, a lot of the time, capacity managers end up fire fighting, leaving the proactive side a far sight. I guess we can blame the initial design / scope phases but back in the day when this tin was laid down, I imagine the tekkies didn't believe demand would spiral out of control the way it often does.

I am now working for a new business in a new field, involving a fresh infrastructure nationwide which will affect every one of us soon enough. It's a godsend to be here because we seem to be adopting an almost textbook ITIL fashion, so I imagine the learning curves will be exceptional! I look forward to sharing more experiences, although I hasten to add I am really a rookie to the whole area so look forward to learning from your experiences!

MSP

[UKVIKING]

The issue with capacity management as i see it is exactly as you stated

The operational IT hand does not know what the business is planning to do.. until it is in place

As I stated, I started doing network capacity mgmt at my last assignment

We tracked the network utilization using eHealth (not a bad tool).
We tracked the network project mgmt team that were 'upgrading' the network b/w

The customer started to want the following from networks

1 - Network device - CPU, Memory utilization and errors
2 - LAN and internal building network utilization

The customer also started to realize that other areas within the business could benefit from some sort of capacity mgmt philosphy

the back up / data storage team started to have their own mtgs

However, the key piece of info was missing... what was the customer planning to do in the next 6 months - more offices, more office (front office ) tools / applications etc

The more tech tool - shiny toy - Video to desktop, video conf, etc etc had a major impact on the network - LAN ->>> WAN.

there was no mechanism / team / group / toolset that could do predicative analysis on what impact the shniny new toys would have

part because Capacity mgmt was basically me. and part because the underlying cmdb concept was weak at best

there was no central repository where I could go..

hmm. the office lan on the 3rd floor is all 10Mb. The switches are .... and need to be upgraded to ... to supprot 100Mb.
the routers/ bridges/ encryption devices are 10/100 Mb capabel but.. what about load ?

etc
etc
etc

Basically, there was not enough info to make info useable

While these issues are noted, there was not a plan to upgrade the processed - config/ service level/ capacity - in a common marching order

once Cap mgmt was in place - as i had done it - the emphasis was still fire fighting

In all honesty, as I dont see a lot of capacity manager's roles out there, I don t see company's worring about cap mgmt as cap mgmt if done properly means expenditure
_________________
John Hardesty
ITSM Manager's Certificate (Red Badge)

Change Management is POWER & CONTROL. /....evil laughter

[MSP]

Viking,

I say this with regret but you're right. Businesses don't seem to adopt a clear cut capacity management process.

However, when your working on a multi-million pound contract that's watched upon by hawks, and service credits are potentially big losses of the businesses' money, then capacity management becomes quite important!

I was formerly looking at network utilisation, interestingly using eHEALTH, also known as Concord along with many other names I might imagine. I know it's a small world so I wonder whether our paths may have crossed!

My new organisation is using the best latest tools, and recognises the importance of every ITIL discipline so I guess that's a good start!

Final point, you're right (again), capacity managers are like gold dust...so it's hard to find opportunities but when you do, the rewards can be quite reasonable!

[UKVIKING]

MSP

I can be found on the Business (IT) professional site that links me to my fellow co-workers

That is a big hint ;=-
_________________
John Hardesty
ITSM Manager's Certificate (Red Badge)

Change Management is POWER & CONTROL. /....evil laughter

[UKIT]

Capacity management is an area that I was asked to get involve with and had to produce a weekly report on server disk space for the operating system and data volumes.
An automated tool was developed that reported all volumes on server's in which the free space was down to 20% or less of capacity.
Disk space and it's management is a continual process and is vital to any business.
Simply adding more disks to a server is not always the answer.

[UKVIKING]

OK

While this tool and the resulting information is gathered, what is being done about the data filling up the space

1 - The old hoary - I got a bigger closet and it is full again

2 - Once you start tracking servers and disk space, you will most likely need to see what is on these servers and what role does the server carry

for example:

SWAP Space vice and memory utilization:
Defragmentation of drives: yes, even ntfs gets fragmentary
Server type and use of the disks:
File servers,
Print servers
mail servers
Web servers
DB Servers
Content servers - video, music etc

Are you using tools that promote storage on network shares
Are you involving multimedia content in the website - to your customers

Then the issue comes -

what about backing up key data
what about storage backed up key data - on site / off site / testing of storage data to verify

how can you identify the users with more crap than other

is your environment set up to have the user use a network share as their My Documents directory

How about Anti virus, anti malware etc on your data

Operational processes
system maintenance
swap space clean up
defragmentation
disaster recovery -- cold warm hot

What happens with the report.
Is it merely reporting or is there a regularly scheduled meeting that discusses and tasks actions

Is there an establish policy on how much space per user, how it is checked, etc

Welcome to capacity management - you have just started with hard drive space

what with Tera, exa and petabyte becomng the scale for storage space and relatively cheap -

Then you have to expand cap mgmt to your network

It does you no favors if you have high capacity storage and usage if the internal network cant support the b/w
_________________
John Hardesty
ITSM Manager's Certificate (Red Badge)

Change Management is POWER & CONTROL. /....evil laughter

[Pascale_it]

Hi,

I recently got involved in ITIL and Capacity Management.
Now, we have some kind of unused metrics and models, and would like to get on control. My company defined some kind of models that have not been maintained, we've got some business KPI(in use), and other business metrics but most of the time not that much accurate (... "these are not wrong estimates but the company that failed to reach the business goals"...). The lack of information and its reliability required a decision: getting a tool, to get information that are accurate (CPU, RAM usage) and avoid negociating metrics, and discussing business estimates.

So here is my question about the use of a AI software integrating ITIL capacity management process and lifecycle (as Neptuny Caplan o Tuppas Capacity tool) and the need of a company to have ITIL certified people.

Isn't it strange to get on these two tracks at the same time? Is there a need of having people "certified" explaining the ITIL model that sustain the tool? Did you use some of these AI software? How does your role as Capacity manager evolved???

If my questions are a bit too opened, do any have some tips about the use of such tools?

Thanks
P.