in this paper common problems and challenges of data center have been identified and methods have been explained to improve the efficiency and reliability of data center
2. In data centers energy cost is one of the major components of the service cost. Service
cost can be reduced by adopting the energy conservation technologies. In eighteen
century, most industrial plants generated their own electricity using different type of fuel.
But when different electrical power plants and national grids were constructed by
government, many industries stopped producing electrical energy and started to
purchase electricity from different electric supply companies. Factors that deprive the
industry from energy conservative technologies were excessive electricity production
and low fuel or energy cost Situation changes completely in nineteen century when
power demand cross the supply limit Fuel cost rises dramatically that causes the
problem of energy crisis. So industries started to think about energy conservative
technologies and they are moving towards complete self-dependent. The technologies
that are energy efficient have become very popular now due to high fuel cost and
uncertainty of fuel supply. For these reasons, data centers in different develop countries
are taking an active role in the increased use of energy conservative technologies.
Today Pakistan is facing the problem of energy crisis and the power supply short fall of
over six thousand megawatts may become worse because the gap between supply
and demand is increasing day by day. This report is designed for rapid introduction of
energy conservative technologies in data centers. This research delves into specific
guidelines on achieving a high level of flexibility and scalability in the data center.
Common challenges in data centers are
Availability
Infrastructure Monitoring
Heat Density
Energy Efficiency
Power Density
Cooling capacity
Space Constraint
1. Equipment of stock :
Stock is not available and not known that what spare part we need in future and
what is the priority of different equipment.
2. A data center infrastructure management (DCIM)
DCIM system is not fully integrated with Power and cooling system. A data center
infrastructure management (DCIM) system provides a single, comprehensive view
into both IT and facility management – two areas previously managed
independently. Implementing a DCIM solution allows data center owners to
efficiently leverage existing synergies.
DCIM includes followings
3. 1. Building management systems
2. Power quality management
3. Fire safety/Security Management
4. Cooling Management
There is no monitoring of such managements
DCIM sector composition
- 55% monitoring (power, environmental)
- 34% Asset, configuration management
- 10% Cooling optimization, environmental management
- 1% Operational CFD analysis
DCIM provides real-time data to understand the true capacity of available infrastructure
Additional monitoring points to the building automation systems to monitor data-center
temperature, humidity, power supply voltage for primary and redundant power sources
and, where utilized, UPS and emergency-generator status through remote monitoring in
the case of a fault. When a parameter goes out of band, a message is sent via remote
paging or e-mail to contact on-call personnel.
3. Cooling system and HVAC
Cooling system and HVAC is not accurately calculated, designed, and understand that
can save up to 32% energy consumption. Most of the staff don’t know even the ABC of
Cooling system
4. Intelligent aisle containment:
The well-established practice of hot / cold aisle alignments sets up another movement—
containment. Aisle containment prevents the mixing of hot and cold air to improve
cooling efficiency and enable higher densities.
Up to 40% of cooling in traditional hot-ailse/cold-aisle doesn't do any work. Air mixes
over the tops of racks and around the rows.
• Hot-aisle/cold-aisle containment uses a physical barrier to separate airflow
through makeshift design solutions, ducted plenum systems and other
commercial offerings. The separation of hot and cold air can provide much better
uniformity of air temperature from the top to the bottom of the rack.
• That uniformity of temperature enables data center pros to raise the set point
temperature more safely. Together with VFDs, containment can reduce fan
energy use by 75%.
5. Lack of Trust in manpower
Manpower are not allowed to touch the electrical equipment. Just trust and have believe
in your manpower and allow them to learn from mistakes and make the necessary
changes to prevent those mistakes in the future. Finally, be courageous about making
changes. It might involve expense and pain, but asking for more money or time is far
better than the potential consequences.
4. 6. Green Data Centers.
Most of the facility mangers have no idea of green data centers. Building and certifying
a green data center or other facility can be expensive, but long-term cost savings can
be realized on operations and maintenance. Another advantage is the fact that green
facilities offer employees a healthy, comfortable work environment. Infrastructure
spread across multiple locations and geographies in Pakistan are a daunting task.
There is a device called Economizers, which use outside air to reduce work, required
by the cooling system, can be an effective approach to lowering energy consumption if
they are properly applied in those areas of Pakistan where weather is cold like quetta,
gilgit etc. Green solutions include:
a. More efficient hardware components and software systems (e.g., systems
virtualization)
b. Innovative cooling systems
c. Using natural ways to cool equipment (evaporation, natural chillers)
d. Building near advantageous natural resources or environments (rivers, cold
climates).
e. Effective server and rack management for better air-flow
7. GSM Alerts:
Critical alarms like fire alarm should be send through GSM on
8. Civil and Electrical Diagrams:
Civil and electrical diagrams (rack diagram, AVR, Generator etc) are not available
9. Commissioning and installation
Commissioning and installation of electrical equipment is done in the absence of
officers. It must be done under the supervision of facility managers so that they
understand the infrastructure basics.
10.Original equipment manufacturer (OEM) vender service
Original equipment manufacturer (OEM) vender service is not called for maintenance
instead of it local vendors are preferred. Vendors generally do a satisfactory job of
getting their own equipment working, and also performing operator training for their own
equipment and sub-systems. However, many field technicians are too focused on
“inside the box” individual component performance, to the point of attempting to prove
specifications that are not relevant to the project.
11.Documentation and reports:
5. Usually after commissioning of electrical equipment there is nothing documented.
Comprehensive commissioning report needs to be developed, including documentation
such as an executive summary, as-tested procedure with results, recorded test data,
issues, resolutions, and recommendations. Although this is a proven, repeatable
process, it needs to be customized per client requirements or in collaboration with other
service providers.
12.Designing approach:
Electrical power is not a donkey or layman work. To prove this we have to think out of
the box and beyond the routine work like laying of cable, starting and stopping of
generator and refilling of fuel etc. facility managers should have designing approach
instead of installation approach.
13.Flexibility in Data Centers:
There must be a research section that work on latest energy efficient technologies
that are flexible enough to hand shake with future technologies and can save huge
amount of energy crisis. Modularity and Flexibility allows data center to grow and
change over time.
14.Alternative energy technologies
The use of alternative energy technologies such as
a. Photovoltaic
b. heat pumps
c. evaporative cooling
d. solar system for lighting and thin clients
e. The use of hybrid or electric company vehicles
15.Optimization of system:
Generators installed at data centers are of large capacity more than needed. This
oversizing of generator causes inefficient and non-redundant operation of power
system. Information and communications technologies contribute 2% of global carbon
emissions, with data centers accounting for 14% of that footprint. At this trend,
Greenhouse Gas (GHG) emissions from data centers are projected to more than double
from 2007 levels by 2020.
16.Lack of ability to view bigger picture of power system:
6. Data center managers are paying more for power and becoming increasingly aware of
new “green” technologies. But nearly 40% of manpower don’t know if their power bill
changed from 2013-2014 – they are not measuring costs or consumption.
17.Cut down the multiple vendor network:
There should be one contractor or vendor that can deal all infra related problems like
a. Power
b. Standby power
c. Cooling
d. Fire protection system
e. DCIM
Typical Data Center Energy Consumption
18.APPLE’S LEED-PLATINUM
data center in Maiden, North
Carolina, built by Holder
Construction, utilizes two 100-
acre solar power farms that
generates 20MW of power
each, as well as a Bloom
Power Conversion,
(11%)
Lighting,
(3%)
Cooling,
(50%)
Servers & Storage,
(26%)
Network Hardware,
(10%)
Equipment
36%
Facility
64%
7. Energy developed biogas-powered power generator (using biogas from nearby
landfills) to add an addition al 4.8MW. Apple estimates that, upon completion, 60%
of the center’s power will be supplied by those on-site renewable-energy power
sources. Apple aims to make it a zero-emission site. Other planned energy efficient
measures include:
a. Chilled water storage system to improve chiller efficiency (for HVAC systems)
b. Precision cool air distribution system
19.Facebook’s LEED-Gold Prineville Data Center in Prineville, Oregon built by DPR
Construction has a
capacity of 28MW. It
uses an innovative
cooling system created
for the unique climate
characteristics of
central Oregon. In
addition, it also utilizes
an evaporative cooling
system and an airside
economizer that brings
in 100% ambient air.
On average, the facility
consumes 28 gallons
per minute, while other
local industrial users consume 60 -173 gallons per minute. Not only does the facility
use energy from local power supplier Pacific Power, it also has built solar arrays to
power office use
8. Another report
1. Poor maintenance:
75% problems in data centers are directly or in directly related to poor maintenance and
ignorance of staff in data center. Performance, maintenance cost, running cost and
efficiency of data center depends upon the maintenance practice. Data center can be
operated with zero percent shut down or no down time by effective and efficient
operation and maintenance programme. Proper maintenance improves customer
service, lower running costs, and freeing up IT staff time to focus on critical business
endeavors. At a typical enterprise data center, about 70 per cent of the IT budget is
spent just on management and maintenance but here we spent 70% of budget on
telephone calls and faulty equipment replacement. This means that very little of your
staff’s time is spent using IT for innovation that will improve your business.
9. 2. Lack of awareness
vendor recommends un feasible solution for our system that not suit and fulfill our
requirement.
3. Limited knowledge of energy efficiency options:
Limited knowledge on energy efficiency options in data centers renders the data center
to operate inefficiently by using old and obsolete technologies. These technologies can
reduce electrical energy consumption or money. It improves service quality, reduce
operational expenses and improve efficiency
4. Short time thinking
Fear of High intial cost. (but my personal experience is that running cost of less efficient
device is more than initial cost while the case is totally reverse in case of more efficient
device)
5. Non-technical lower and upper staff
6. No research on latest technology
There should be research department or section.
7. Nonstandard local venders
One way to improve your business-critical application availability is with a primary
service vendor who has expertise in that field. You might start by using the authorized
vendor to provide a one-time evaluation of your current environment, and then
recommend improvements.
8. Non learning environment
There is a need to increase the staff skill and their training, needed to keep abreast of
this convoluted world.
9. International standards are not followed for electrical system
IEEE, IEC, NEC and other standards must be followed
10. Dealing with multiple vendors
A range of service levels provides high-quality technical assistance and operational
excellence – and can condense the complexity of a multi-vendor environment down to a
single point of accountability.
11. Oversizing of generator is highly inefficient
There is a misconception about oversizing. Generator efficiency is maximum
when generator operates at 100% load and decreases with the decrease in
running load.
10. 12. Oversizing of cooling is highly inefficient
13. Designing draw backs
14. Voltage fluctuation in utility power supply
15. Unnecessary SLAs and other agreements
16. No symmetry in power and cooling system design
Some data center uses room based cooling, some row based and some rack
based in short there is no symmetry.
17. Monopoly of one or two non-standard, non-technical, quack vendors.
18. No check and balance on electrical device’s prices given by vendors.
19. Total dependency on vendors
20. Impersonated by vendors as if our data center is tier 3 type.
21. No inventory of power and cooling system or in the hand of local
vendors that sell it at high rates.
22. Poor Power management system
Power management system successfully achieve the Uptime Institute’s Tier Four level
of 99.995 per cent uptime (which translates to just 24 minutes of unplanned downtime
per year), the power management issue can prove costly even in traditional data
centers. When a server uses more power than it was intended to consume, it threatens
the availability of every application it supports. Similarly, if a rack is drawing too much
power, it puts every device connected to it at risk of outage should it fail. Without a way
to monitor when a rack or other equipment is nearing its power capacity, IT has no way
to predict and prevent such failures. Similarly, faulty devices may create a power surge
on the network, resulting in tripped circuit breakers that can bring down multiple servers.
Here again, with no way to tell what the offending device is, IT is powerless to prevent
such an incident.
23. Poor air flow management system
Cooling is not equally distributed in case of row based cooling
24. Confused between oversizing and redundancy:
25. Hot and cold aisle containment is absent
26. No blanket plates are installed in used rack space.
27. Hot and cold aisle containment is not totally closed
Maintaining strong separation between hot and cold aisles is particularly important.
Even when implementing hot aisle and cold aisle separation, about 40% of the cold air
short circuits past the equipment it is supposed to cool. Blocking off open gaps in
cabinets with blanking plates is key as is fitting empty (and fully blanked) cabinets into
gaps or alternatively filling full-length curtains in gaps.
11. 28. No measure and inspection of power and cooling system on daily
basis.
Record and track power use regularly. Benchmark your existing energy usage against a
comparable timeframe. Continue to measure energy usage consistently and regularly.
Regular meter readings will help you better understand trends and eliminate any
seasonal spikes or oddities. At the end of a 12-month period you should have enough
data to benchmark against and get a much clearer picture of your overall efficiency.
Cutting energy usage will help organizations save significant money, while changing the
public perception of data centers being “energy hogs” with a “greener” impression. To
reduce energy consumption, it must first be accurately measured. The measurement
must start with a baseline view of what energy every part of the data center is actually
consuming. And to be accurate, this view must include all areas such as IT equipment,
power distribution infrastructure, and ventilation/cooling infrastructure. This is an often
time-consuming and manual data extraction process to which nobody looks forward.
Poor air flow management system causes hotspots in data center. Block up all holes in
the ceiling, walls or any other aperture such as old air conditioning ducts that will allow
air to escape or unconditional air to enter the data center. The voltage of your electricity
transformers should be checked regularly to ensure they match exactly the supply
voltage requirement for your equipment. A higher voltage than required means an
unnecessary use of power, increasing costs.