Green cloud computing

1,920 views
1,802 views

Published on

Published in: Technology
0 Comments
1 Like
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total views
1,920
On SlideShare
0
From Embeds
0
Number of Embeds
39
Actions
Shares
0
Downloads
220
Comments
0
Likes
1
Embeds 0
No embeds

No notes for slide
  • It denotes the infrastructure as a “Cloud” from which businesses and users can access applications as services from anywhere in the world on demand. Many computing service providers including Google, Microsoft, Amazon, and IBM are rapidly deploying data centers in various locations around the world to deliver Cloud computing services. Cloud computing delivers infrastructure, platform, and software (applications) as services, which are made available to consumers as subscription-based services under the pay-as-you-go model. Cloud computing offers significant benefits to IT companies by freeing them from the low-level task of setting up basic hardware and software infrastructures and thus enabling focus on innovation and creating business value for their services
  • number of users who log on to the Australian Open web page. The spikes correspond to the month of January during which the tournament is going on. It would be wasteful to have servers which can cater to the maximum need, as they won’t be needed during the rest of the year.
  • The iPad relies upon cloud-based computing to stream video, download music and books, and fetch email. Already, millions access the ‘cloud’ to make use of online social networks, watch streaming video, check email and create documents, and store thousands of digital photos online on popular web-hosted sites like Flickr and Picasa The term cloud, or cloud computing, used as a metaphor for the internet, is based on an infrastructure and business model whereby - rather than being stored on your own device - data, entertainment, news and other products and services are delivered to your device, in real time, from the internet. The creation of the cloud has been a boon both to the companies hosting it and to consumers who now need nothing but a personal computer and internet access to fulfill most of their computing needs.
  • The goals are similar to green chemistry; reduce the use of hazardous materials, maximize the energy efficiency during the product’s lifetime, and promote recyclability or biodegradability of defunct products and factory waste.
  • Software as a Service: However at a particular time a particular number of instances of the software are allowed to be running per user. This feature protects the software from piracy and illegal access. One example of such kind of software is Google Docs. Storage as a Service: In this case the user PC with a solid state drive is sufficient as cloud acts as primary storage device. Files which are present on cloud can be accessed from any computer connected to internet this give mobility to user data.
  • Now we will analyze the percentage of total power consumption in transport, storage and servers/computation, as a function of downloads per hour, for a public cloud and private cloud storage service. Case 1: At low download rate (10 -2 /h) Approximately Public Cloud : 75% of power consumed in storage 25% is consumed in transport And the remainder is consumed by servers Private Cloud : 90% of power consumed in storage 10% is consumed in transport And the remainder is consumed by servers Case 2: At more than one download per hour Approximately Public Cloud: 10% of power consumed by server Storage less than 1% And the remainder is consumed in transport Private Cloud: 35% of power consumed by server Storage less than 7% And the remainder is consumed in transport Hence we deduce: Case 1: The power consumption in storage dominated the total power consumption for both public and private cloud storage services at low usage levels. Case 2: As the average download rate increases, we need an increased number of servers, routers and switches to support the additional traffic. Hence the power consumed in server and transport increases, while the percentage of total power consumed in storage decreases.
  • The power consumption of the storage service is below 1W at low download rate (<one download per hour per file). As the download rate increases, due primarily to the increased power consumption in transport, the power consumption of the storage services increases towards 10W. The power consumption of the public cloud storage is service is about 2.5 times that of the power consumption of the private cloud storage service, at medium and high download rates, due primarily to the increased energy consumption in transport. Now when we compare the energy consumed by a cloud storage service to a HDD (2.5” HDD) in home computer that is idle (low power state) 75% of the time and active 25% of the time. We can clearly say that at low download rates, the storage service is more efficient, but this benefit vanishes if the number of regularly used files is larger, and if downloaded frequently.
  • Each user has a monitor running at a resolution of 1280 X 1024 with 24-b color, giving a total 1280 X 1024 X 24 b/frame. If Y is the number of new frames every second (frames/s), the data rate between each user and the server is A = 1280 X 1024 X 24 Y b/s. We model a data center with computation servers and consider two scenarios. In the first scenario, each server is able to support 20 users and in the second scenario each server is able to support 200 users. The utilization of the Internet is 50% and the utilization of the corporate network is 33%, which is sufficiently low to minimize latency. For frame rate below 10 -2 Public Cloud: Less than 10% is consumed in transport   As the frame rate increases, the percentage of power consumed in transport significantly increases.   Private Cloud: For frame rate below 10 -2 With 200 users per server, transport, storage and servers together consumes less than half of total power consumption. The remaining power is consumed in the terminal.     With 20 users per server, at frame rates less than 0.1frame/s, transport consumes less than 5% of the total power, increasing to 40% of total power consumption as frame rates increase to 1 frame equivalents per second.   With 20 users per server, the majority of power is consumed in the servers.  
  • The power consumption of the cloud services with 20 users per server is 35–45 W when the frame rate is small (G 0.1 frames/s). If the transport component of the public cloud service is required to support the equivalent of 1 frame/s, the power consumption of the service rises to 129 W due to the high transport requirements. The power consumption of the private cloud service with 20 users per server does not exceed 60 W even at high frame rates. The power consumption of the cloud services with 200 users per server is 12–23 W. The power consumption in transport increases as the frame rate increases, but the transmission rate limit of each server of 800 Mb/s limits the frame rate to 0.11frames/s. In the above figure we also have power consumption at idle of a low-end laptop (18 W) and the power consumption at idle of a modern midrange computer (70 W). A low-end laptop consumes the least power but also has the least functionality and processing capacity. The cloud service in both scenarios is more efficient than the modern mid-range PC at low frame rates. However, as the frame rate increases the power consumption of the public cloud service with 20 users per server approaches and then exceeds the power consumption of the midrange PC. Cloud software services are more efficient than modern midrange PCs for simple office tasks, where the number of users per server can be high. However, if the user’s tasks are intensive and high frame rates are required, then public software services is not energy efficient relative to a modern midrange PC.
  • At fewer than 10 -1 encodings per week over 90% of power is consumed in the user’s laptop for both the public and private cloud processing services. As the number of such encoding per week increases, the energy consumption in transport and processing increases. The user’s laptop is modeled as being used for 40 h/week regardless of the number of encodings and so its energy consumption remains constant as the number of encodings increases. At one encoding per week with a public cloud processing service, approximately 40% of total energy is consumed in servers, approximately 15% of total energy is consumed in transport, and the remainder is consumed in the user laptop. For a private cloud processing service with one encoding per week, half of the total energy is consumed in the user laptop, approximately 50% of total energy is consumed in servers, and the remainder is consumed in transport. The trend of increased energy consumption in servers and transport continues as the number of encodings per week increases. The results indicate that in a public cloud processing service, even for the computationally intensive task of video encoding, transport consumes a significant percentage of total energy consumption at medium and high usage rates. However, the percentage of energy consumed in transport with a private cloud processing service is less than 5% at all usage rates
  • At less than 10 -1 encodings per week, total energy consumption with the public and private cloud processing services is similar. If on average one encoding is performed each week, the total energy consumption with the public processing service is 1.6kWh/week. At the same number of encodings per week, the total energy consumption of the private processing service is 13% lower at 1.4kWh/week due to the lower energy consumption in transport. The total energy consumption of the public cloud processing service increases to 10 kWh/week at ten encodings per week and approaches 100 kWh at the extreme of 100 encodings per week. The private cloud processing service is approximately 21% lower, again due to the lower energy consumption in transport. to encode a typical DVD video file requires 13.2 h on the old midrange PC, 4 h on the modern midrange PC, and 2.2 h on the high-end PC. Therefore, a maximum of 9.7, 32, and 58 encodings can be performed on the old midrange PC, modern midrange PC, and high-end PC, respectively. Performing 40h of common office tasks consumes 5, 2.8, and 5.6 kWh/week on the old midrange PC, modern midrange PC, and high-end PC, respectively. Thus, users with older generation computing equipment could achieve significant energy savings as well as increased computational capability by moving to a combination of a modern low-end laptop and cloud computing.
  • Green cloud computing

    1. 1. GREEN CLOUD COMPUTING Rohit Sinha (2BV07CS081) www.rohitsinha.com BVBCET - Hubli
    2. 2. CLOUD COMPUTING ?• Cloud computing is a model of enabling convenient, on-demand network access to shared pool of configurable computing resources that can be rapidly provisioned and released with minimal management effort or service provider interaction.
    3. 3. NEED FOR CLOUD COMPUTING January Month
    4. 4. Happening Cloud Computing• The iPad relies upon cloud-based computing to stream video, download music and books, and fetch email.• Google’s signature products - Gmail, Google Documents and Google Earth - are delivered from the cloud.• Its ambitious project to create a digital library will be entirely hosted by servers storing most of the world’s published work, all in digitized form.
    5. 5. GREEN CLOUD COMPUTING• Green computing is the study and practice of using computing resources effectively.• Green Cloud computing is envisioned to achieve not only efficient processing and utilization of computing infrastructure, but also minimize energy consumption.
    6. 6. CLOUD SERVICE MODELS• Software as a Service: The software is present on clouds and all type processing is done on cloud only. This feature allows users to access the software from any computer which is connected to the internet.• Storage as a Service: This service allow user to outsource their data storage needs to the cloud. The user store all his/her files on cloud but the all kind of processing is done on user’s PC• Processing as a Service: This model provides user the functionality to process complex computation on cloud which consists of powerful servers. Tasks which demand less processing power are carried on user’s PC only.
    7. 7. SUMMARY OF MODELS   Software as a Service Storage as a Service Processing as a Service Short task at client,Location of Processing Cloud Client Large tasks in cloud Location of Storage Cloud Cloud Client Transmit commands andFunction of Transport All files/documents Files for large tasks receive results
    8. 8. ANALYSIS OF CLOUD SERVICES• In this section we will compare the per-user energy consumption of each cloud service mentioned in section II using energy models. We will also compare the energy consumption against the energy consumption of conventional computing.• We have two types of cloud public cloud and private cloud. In this section ET is the per bit energy consumption of transport. For• Private cloud: ET = EC which represent transport through corporate network.• Public cloud: ET = EI which represent transport through internet.
    9. 9. STORAGE AS A SERVICEPUBLIC CLOUD PRIVATE CLOUD
    10. 10. Per user power consumption of public and private cloud storageservices as function of download rate. Also included is the power consumption of a modern laptop HDD. The average document size is 1.25MB
    11. 11. SOFTWARE AS A SERVICEPUBLIC CLOUD PRIVATE CLOUD
    12. 12. Per-user power consumption of public and private cloud software as afunction of download rate. Also included is the power consumption of a low- end laptop and the power consumption at idle of a modern midrange computer
    13. 13. PROCESSING AS A SERVICEPUBLIC CLOUD PRIVATE CLOUD
    14. 14. Power consumption of public and private cloud processing services as afunction of encoding per week. Also included is the power consumption of amodern midrange PC and the power consumption of a modern idle of a high- end PC
    15. 15. SUMMARY OF RESULTS Software  as  a  Processing  as  a Energy Component Service Type Storage as a Service Service Service Medium to high Public High frame rates Always encoding per weekTransport Private Never High download rates Never Public Never Low download rates -Storage Private Never Low download rates - Medium to high Public Few users per server Never encoding per weekProcessing Medium to high Private Few users per server High download rates encoding per week
    16. 16. CONCLUSION• Power consumption in transport represents a significant proportion of total power consumption for cloud storage services at medium and high usage rates.• For typical networks used to deliver cloud services today, public cloud storage can consume of the order of three to four times more power than private cloud storage due to the increased energy consumption in transport.• Nevertheless, private and public cloud storage services are more energy efficient than storage on local hard disk drives when files are only occasionally accessed. However, as the number of file downloads per hour increases, the energy consumption in transport grows and storage as a service consumes more power than storage on local hard disk drives.• In cloud software services, power consumption in transport is negligibly small at very low screen refresh rates. As a result, cloud services are more efficient than modern midrange PCs for simple office tasks.• At moderate and high screen refresh rates, power consumption in transport becomes significant and energy savings over midrange PCs are reduced.• Significant energy savings are achieved by using low-end laptops for routine tasks and cloud processing services for computationally intensive tasks, instead of a midrange or high-end PC, provided the number of computationally intensive tasks is small.
    17. 17. BIBLOGRAPHY• [1] CISCO. (2009). CISCO VISUAL NETWORKING INDEX: FORECAST AND METHODOLOGY, 2009–2014. WHITE PAPER. [ONLINE]. AVAILABLE: HTTP://WWW.CISCO.COM.• [2] A. WEISS, “COMPUTING IN THE CLOUDS,”NETWORKER, VOL. 11, NO. 4, PP. 16– 25, 2007.• [3] B. HAYES, “CLOUD COMPUTING,” COMMUN. ACM, VOL. 51, NO. 7, PP. 9–11, 2008.• [4] T. SINGH AND P. K. VARA, BSMART METERING THE CLOUDS,[ IN PROC. IEEE INT. WORKSHOPS ENABLING TECHNOL., INFRASTRUCTURES FOR COLLABORATIVE ENTERPRISES, GRONINGEN, THE NETHERLANDS, JUN.–JUL. 2009, PP. 66–71.• [5] D. KONDO, B. JAVADI, P. MALECOT, F. CAPPELLO, AND D. P. ANDERSON, “COST- BENEFIT ANALYSIS OF CLOUD COMPUTING VERSUS DESKTOP GRIDS,” IN PROC. IEEE INT. SYMP. PARALLEL DISTRIB. PROCESS., ROME, ITALY, MAY 2009, DOI: 10.1109/IPDPS. 2009.5160911.• [6] OPEN CLOUD MANIFESTO. [ONLINE]. AVAILABLE: HTTP://WWW.OPENCLOUDMANIFESTO.ORG/• [7] GOOGLE DOCS. [ONLINE]. AVAILABLE: HTTP://DOCS.GOOGLE.COM• [8] AMAZON WEB SERVICES. [ONLINE]. AVAILABLE: HTTP://AWS.AMAZON.COM• [9] AZURE SERVICES PLATFORM. [ONLINE]. AVAILABLE:
    18. 18. THANK YOU

    ×