The document discusses strategies for upgrading aging data centers to extend their lifecycles. It outlines key drivers for upgrading like capacity, resilience, and efficiency. It recommends developing a thorough understanding of the existing facility through inspection and documentation review to understand constraints, hidden issues, and actual capacities. A multi-step masterplan is suggested to upgrade systems iteratively over time through modular additions rather than replacing everything at once. Integrating mechanical, electrical, and IT systems is also advised.
1. EXTENDING THE LIFECYCLE
OF AGEING DATA CENTRES:
HOW AND WHAT
SHOULD DC’s DO
FOR RETROFITS
Jonathan Price
4th August 2015
2. KEY DRIVERS FOR UPGRADING A DATA
CENTRE
Capacity
Resilience
Efficiency
Lifecycle
Replacement
Combina5on
of
the
Above
All
images
and
copyright
belong
to
original
owner
and
are
reproduced
here
for
the
purposes
of
training
and
educa7on
only
3. WHAT ARE THE BUSINESS NEEDS ?
All
images
and
copyright
belong
to
original
owner
and
are
reproduced
here
for
the
purposes
of
training
and
educa7on
only
Initial
upgrade to
meet
immediate
business
needs
Iterative
planned
upgrade
steps
Final
planned
upgrade to
deliver final
capacity
End of
lease / New
DC
available
Capacity
Resilience
Standards
4. Market Drivers
All
images
and
copyright
belong
to
original
owner
and
are
reproduced
here
for
the
purposes
of
training
and
educa7on
only
Efficiency
Wider
Temperature
Bands
Resilience
Changes
in
Guidelines
Cloud
Services
Dynamic
Load
Fast
Deployment
Standards
Compliance
Constantly
Changing
Data
Centre
Use
Co-‐loca>on
Enterprise
5. DO YOU REALLY NEED TO UPGRADE AND
WHEN?
All
images
and
copyright
belong
to
original
owner
and
are
reproduced
here
for
the
purposes
of
training
and
educa7on
only
6. UNDERSTANDING THE CONSTRAINTS
» Timescales
» Possibility of Outages
» Level of existing information
All
images
and
copyright
belong
to
original
owner
and
are
reproduced
here
for
the
purposes
of
training
and
educa7on
only
7. THE OBVIOUS CHALLENGES
All
images
and
copyright
belong
to
original
owner
and
are
reproduced
here
for
the
purposes
of
training
and
educa7on
only
Paralleled Systems
Lack of Isolation Points
Space Constraints
Capacities of Utilities
9. SOME HIDDEN ISSUES
Documentation
TechnicalCommissioning
All
images
and
copyright
belong
to
original
owner
and
are
reproduced
here
for
the
purposes
of
training
and
educa7on
only
Commissioning
» Existing systems not meeting
design capacity
» In-operable isolation points
» Unresolved defects
10. SOME HIDDEN ISSUES
Documentation
TechnicalCommissioning
All
images
and
copyright
belong
to
original
owner
and
are
reproduced
here
for
the
purposes
of
training
and
educa7on
only
Technical
» Neutrals
» Circuit breaker discrimination
» Surge Suppression
» Changes in Code &
Regulations
» Latent defects
» Obsolete equipment
» Equipment fed from multiple
sources
11. SOME HIDDEN ISSUES
Documentation
TechnicalCommissioning
All
images
and
copyright
belong
to
original
owner
and
are
reproduced
here
for
the
purposes
of
training
and
educa7on
only
All of the Above
» Control systems
» Power supplies to ancillary
equipment
» Non-monitored points
leading to potential latent
defects
» Single cord equipment
» IT load distribution
12. KNOWLEDGE, KNOWLEDGE AND MORE
KNOWLEDGE
» Understanding of existing facility is vital
» Detailed site inspection
» Discussion with facilities managers:
» Evidence of existing facility not meeting
design capacities
» Know equipment defects
» Commissioning data
» Existing operational challenges
(unsyncronised supplies from different
sources)
» Maintenance regimes (more risk to process if
switching / operating equipment that has not
been maintained – supportability of
equipment / spares available on site
All
images
and
copyright
belong
to
original
owner
and
are
reproduced
here
for
the
purposes
of
training
and
educa7on
only
13. KNOWLEDGE, KNOWLEDGE AND MORE
KNOWLEDGE
» Actual capacities based on all constraints (kW vs. kVA
etc.)
» Verify meter readings – only analogue meter may not be
giving you accurate info
» Where documentation is not there – contractor
investigation or testing to verify configuration /
performance
» SPOFs as a result of distribution of IT load
All
images
and
copyright
belong
to
original
owner
and
are
reproduced
here
for
the
purposes
of
training
and
educa7on
only
14. DO YOU NEED TO UPGRADE THE WHOLE PLANT
AT ONCE ?
» Short term, do you need to upgrade at all – retro-
commission
» Is there as simple way to utilise stranded capacity.
» Due to standard equipment sizes not all equipment
needs to be upgraded at once
» Develop masterplan
» Each step should inform the previous (e.g..
mechanical plant upgrade may increase efficiency
or power factor enough to delay or even negate the
need for a mains upgrade)
All
images
and
copyright
belong
to
original
owner
and
are
reproduced
here
for
the
purposes
of
training
and
educa7on
only
15. MASTERPLAN DEVELOPMENT
All
images
and
copyright
belong
to
original
owner
and
are
reproduced
here
for
the
purposes
of
training
and
educa7on
only
IT Load 550 560 570 580 590 600 610 620 630 640 650 660 670 680 690 700 710 720 730 740
Existing
UPS
New UPS
Generator
Mains
Switchgear
Cooling
16. WHICH TECHNOLOGIES WILL WORK ?
All
images
and
copyright
belong
to
original
owner
and
are
reproduced
here
for
the
purposes
of
training
and
educa7on
only
17. USE OF MODULAR PLANT AND CONSTRUCTION
» Provide capacity as required – delay capital spend
» Higher efficiency due to higher plant loading – could
delay upstream infrastructure upgrade
» May reduce space requirements
All
images
and
copyright
belong
to
original
owner
and
are
reproduced
here
for
the
purposes
of
training
and
educa7on
only
18. CONSIDER SITE AS A WHOLE – MECHANICAL
AND ELECTRICAL INTEGRATION
All
images
and
copyright
belong
to
original
owner
and
are
reproduced
here
for
the
purposes
of
training
and
educa7on
only
» Can be more cost effective to add new independent
mechanical systems instead of augmenting an existing system
» Cost of controls resilience and expansion upgrades
» Cost of electrical resilience and expansion upgrades
» Independent plant used for transition
» Transition risk reduced on existing infrastructure
» New plant probably more efficient
» Mechanical plant replacement can improve power factor and
may negate electrical upgrades
» Consider electrical equipment with reduced cooling
requirements
19. IT AND FACILITIES INTEGRATION
» Avoid the management derived SPOFs
» Consider the cloud environment
Source en.wikipedia.org/wiki/Single_point_of_failure
All
images
and
copyright
belong
to
original
owner
and
are
reproduced
here
for
the
purposes
of
training
and
educa7on
only
20. IT AND FACILITIES INTEGRATION
Source en.wikipedia.org/wiki/Single_point_of_failure
All
images
and
copyright
belong
to
original
owner
and
are
reproduced
here
for
the
purposes
of
training
and
educa7on
only
» Avoid the management derived SPOFs
» Consider the cloud environment
» Consider the use of DCIM
21. DESIGN AND IMPLEMENTATION
All
images
and
copyright
belong
to
original
owner
and
are
reproduced
here
for
the
purposes
of
training
and
educa7on
only
» Can you have an outage?
» Potential significant cost saving
» Can significantly de-risk the process but beware of
hidden risks
» Detailed transition plan is vital?
» Transition plan will generally drive the design
» If outage possible – transition plan just as important to
minimise outages
» Transition plan revisited throughout the project
» Can new permanent plant be used in lieu of temporary plant?
22. DON’T MAKE THE SAME MISTAKES AGAIN
» Discrimination study and match switchgear
» Provide connection points for future – try to make them
technology agnostic
» Be vendor neutral where possible
» Consider adding temporary generator connection points
and possibly alternative supplies to DBs feeding single
cord loads
» Avoid paralleled systems where possible – future
obsolescence and vendor lock in
» Monitor and monitor some more – IT load as well as
infrastructure
» Commission the site properly
All
images
and
copyright
belong
to
original
owner
and
are
reproduced
here
for
the
purposes
of
training
and
educa7on
only
23. CASE STUDY – INITIAL CONFIGURATION
All
images
and
copyright
belong
to
original
owner
and
are
reproduced
here
for
the
purposes
of
training
and
educa7on
only
24. CASE STUDY – STEP 1
All
images
and
copyright
belong
to
original
owner
and
are
reproduced
here
for
the
purposes
of
training
and
educa7on
only
New generator system required
for capacity utilised for
transition
New chillers provided for
capacity can be used in lieu of
existing chillers for transition
On-floor cooling units provided
with diverse supply from new
generator system
25. CASE STUDY – STEP 2
All
images
and
copyright
belong
to
original
owner
and
are
reproduced
here
for
the
purposes
of
training
and
educa7on
only
New redundant
UPS installed –
initially only fed
from the new
generators
A and B static transfer
switches installed – one at a
time so no outage to IT
equipment
26. CASE STUDY – STEP 2A
All
images
and
copyright
belong
to
original
owner
and
are
reproduced
here
for
the
purposes
of
training
and
educa7on
only
Existing MSB and
Generator can now
be isolated
27. CASE STUDY – STEP 3
AT
S
ATS
New MSB installed with
increased capacity
Old Generator System integrated
with the new one
ATSs installed on the existing chiller
DB and existing UPS Input DB and
generator feed removed from
existing MSB
Up-rated mains DB for on-floor
cooling installed upstream of
existing board. Redundant unit from
existing N+1 UPS system used for
back-up
All
images
and
copyright
belong
to
original
owner
and
are
reproduced
here
for
the
purposes
of
training
and
educa7on
only
From
Gen
From
Gen
28. CASE STUDY – STEP 4
All
images
and
copyright
belong
to
original
owner
and
are
reproduced
here
for
the
purposes
of
training
and
educa7on
only
New UPS Installed for Capacity
From
Gen
ATS
New
on-‐floor
cooling
DB
ATS
From
Gen
29. All
images
and
copyright
belong
to
original
owner
and
are
reproduced
here
for
the
purposes
of
training
and
educa7on
only
Jonathan
Price,
Director
–
Mission
Cri>cal
Norman
Disney
&
Young
j.price@ndy.com
+61
2
9928
6868
+61
424
969
015