0% found this document useful (0 votes)
17 views4 pages

GC 2

The document discusses sustainability and green metrics for measuring the environmental impact of IT infrastructure and data centers. It covers topics like sustainability dimensions, a methodology for integrating sustainability into the software development lifecycle, metrics for measuring various sustainability aspects of software and infrastructure, and green metrics for measuring the environmental performance of data centers.

Uploaded by

Mithun B M
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
17 views4 pages

GC 2

The document discusses sustainability and green metrics for measuring the environmental impact of IT infrastructure and data centers. It covers topics like sustainability dimensions, a methodology for integrating sustainability into the software development lifecycle, metrics for measuring various sustainability aspects of software and infrastructure, and green metrics for measuring the environmental performance of data centers.

Uploaded by

Mithun B M
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

Green normally refers to environmental aspects, and sustainability encompasses

environmental, economic and social dimensions. These three dimensions are integrated
and interrelated: • Economic growth without addressing its environmental impact
diminishes quality of life. • Environmental protection without considering local culture and
traditions creates social injustice. • Social disruptions like riots and war jeopardize the
environment and the economy. • A healthy economy enables a better environment and
social welfare.

This methodology is integrated at the end of the existing software life cycle when the
product is released and it includes three steps: 1. Metrics are collected at the end of
development cycle. 2. Metrics are analysed based on their environmental, social and
economic benefits. 3. Corrective or refined actions are formulated for the next development
cycle and release of the product. (Diagram - Development, testing, Release, Requirements)
(Collect metrics, analyze results, define metrics)

Modifiability - Distance from Main Sequence, Reusability - Instability and Abstractness


Portability - Estimated System Lifetime, Supportability - Support Rate, Performance -
Relative Response Time, Dependability - Defect Density Testing Efficiency and Testing
Effectiveness, Usability - Learnability, Effectiveness - Error Rate and Satisfaction
Accessibility - Accessibility Requirements Score, Predictability - Estimation Quality Rate
Efficiency - Project Efficiency, Carbon footprint - Number of Offsite Meetings, Long-Haul
Roundtrips and others (depends on the nature of the project)

Code Metrics tools - Java: – IBM Rational Software (commercial): – Eclipse Metrics (open
source): NET: – NDepend • C++: – CppDepend (commercial): Adobe Flex: – FlexMetrics
(open source): Metrics – ItDepends (open source): • Multilanguage: – Sonar (open source
key elements of IT infrastructure of a data center: • Server design and server systems
development in support of efficient data center service provision and range of service
function. • role of networking within a data center. • role of storage and types of storage
provision. •changing shapes of data center IT platforms through system innovation

Servers - Rack-mounted servers, blade servers, and containers. Network - LAN, WAN,
Routers. Storage - NAS, NFS, SAN

IT platform innovation - Server Farm (Cluster) computing is characterized by multiple,


physically discrete machines, closely linked to provide logical interface of a single machine.
Grid Computing - The core aim of grid computing was to integrate disparate resources
across organizational domains into what became termed virtual organizations.
Service Orientation Service-oriented architecture (SOA), of which Web service is one
instantiation, promotes separation of concerns between service implementation (software)
and service hosting (server hardware). It is a means of providing both data and processing
resources over a network that decouples the service instantiation from a machine-readable
service interface
Virtualization - Rather than a single hardware machine supporting a single OS, that in turn
hosts a single server application (Web server, mail server, etc.), a virtualized system enables
a single hardware machine running a single OS to host multiple virtual machines, which
may or may not be running the same OS. Cloud Computing

Two main functions of the supporting infrastructure are as follows: 1. Ensure that the IT
and facilities are supplied with power at all times. 2. Keep the data center at the required
temperature by removing the heat generated.
AC power is supplied to the main utility station and routed via switch gear to the
substation supporting the data center. Under normal operation the UPS acts as a filter
smoothing the incoming AC and routing to the PDUs whilst at the same time the AC power
charges the DC batteries. In an electrical outage the battery DC power is converted to AC
and routed via the PDUs to the IT equipment racks until such time as the backup generators
can come online to support the IT load.
AC versus DC Power Given the inefficiency of the AC/DC conversion process prominent in
UPS supported data centers and the fact that IT components ultimately utilize DC power,
there have been renewed calls to move data centers to a DC-based power infrastructure.
The theoretical case for DC power encounters little argument as the reduced number of
conversions (i.e. one), where facility supply to the data center is AC, results in reduced
conversion losses and hence greater efficiency.

Cooling The objective of cooling is to ensure that components within the IT equipment do
not overheat, causing damage or degrading the performance and thus impacting service.
The most common means of cooling is by means of convection, which is the transfer of
energy between an object and its environment, due to circular motion of a fluid (e.g. water)
or a gas (e.g. air). Some emerging cooling solutions utilize direct touch – conduction – to
remove heat from servers and IT equipment. Three types of air-based systems:
Room-based, Row-based and Rack-based.
‘Row-based’ distribution is more efficient than ‘room-based’ systems, due to shortened air
paths, whilst ‘rack based’ systems are the most efficient given the reduced power required
to move air within the confines of the rack itself. All three employ a computer room air
conditioner (CRAC) or computer room air handler (CRAH). Under usual operation, heat
within the air is transferred via the CRAC or CRAH at the room, row or rack level, to a liquid
medium, normally water, chilled water or glycol. This heat is subsequently transported
externally and dissipated to the outside air. Wet-side or water-side economization is one
best practice employed to reduce energy consumption by minimizing reliance on costly
mechanical refrigeration.

Server power mgmt control loop


Consolidated data centre server power management usage model - 1. Perform
real-time server power monitoring, Reduce stranded power by scheduling available data
centre power to actual server power consumption, Real-time monitoring of power
consumption Manage data centre hot spots Power and thermal scheduling Power trending
and forecasting. 2. Power guard rail: Impose power guard to prevent server power
consuming beyond a pre-set limit, Deterministic power limit and guaranteed server power
consumption ceiling, Maximize server count per rack and therefore CapEx return on
investment (ROI) per available rack power when rack is under power budget with
negligible per server performance impact. 3. Static power capping: Operate servers under a
permanent power-capped regime, Operate under impaired power availability conditions,
Maximize per rack performance yield when rack under power budget Application power
optimization Application performance compensation.

Virtualization vs traditional data centres - Virtualized cloud data centers introduce


additional degrees of freedom that are not available under traditional operating models
where there is a hard binding between servers and the applications which they run. Cloud
applications run in virtualized environments allowing the dynamic consolidation of
workloads. Firstly, applications hosted in virtualized cloud data centers run on virtualized
OSs, that is, the OS does not run on the bare metal but is mediated through a virtualization
hypervisor. The practical effect is that applications are no longer bound to a physical host
and can be moved around within a pool of servers to optimize the overall power and
thermal performance of the pool. ∙ Secondly, the loose binding between applications and
hosts allows treating a group of hosts as pooled resources, allowing optimizations as a
group that were not possible with individual machines, such as powering down some
equipment during low demand.

Green Data Centre Metrics - Power usage effectiveness (PUE) and its reciprocal data
center infrastructure efficiency (DCiE) PUE is calculated by dividing total facility power (i.e.
all power-consuming elements that constitute the data center eco-system) by IT equipment
power. The metric indicates how effectively or energy efficiently one is supporting the IT
load. Ideally PUE would equal 1.

CUE- CO2 emissions caused by the total data centre energy and IT equipment energy or
CEF × PU
WUE: A data center manager divides the annual site water usage in liters (L) by the IT
equipment energy usage in kilowatt hours (Kwh). Water usage includes water used for
cooling, regulating humidity and producing electricity on-site.
WUEsource , a source-based metric that includes water used on-site and water used
off-site in the production of the energy used on-site.

You might also like