Lecture 55 Sustainable Cloud Computing - II

Summary

Highlights

• The exponential rise in computational needs drives the need for sustainable cloud computing 🌟.
• Efficient data center management can significantly reduce the carbon footprint of cloud computing 💨.
• Intelligent power management and the use of renewable energy are key to sustainable cloud infrastructures 🌞.
• Increasing server density challenges cloud computing sustainability due to higher energy and cooling needs 🔥.
• Utilizing waste heat can contribute to more efficient energy use in data centers ✨.

Key Takeaways

• Sustainability in cloud computing is crucial due to the exponential increase in computational power demands 🌍.
• Data centers heavily impact sustainability due to their significant energy consumption and heat production 🔋.
• Innovative cooling techniques and waste heat utilization can enhance energy efficiency in data centers ❄️.
• Virtualization and thermal-aware scheduling are vital components for optimizing cloud infrastructure sustainability 🌬️.
• Cloud computing sustainability encompasses power management, renewable energy usage, and efficient capacity planning 🔄.

Overview

Chapters

• 00:00 - 01:00: Introduction to Sustainable Cloud Computing In this introductory chapter on sustainable cloud computing, the discussion builds upon previous classes that covered the basics of cloud computing. The focus is on understanding the principles of sustainable cloud computing, which aims to minimize the environmental impact of cloud service provision. Key points include reducing energy consumption, optimizing resource usage, and incorporating environmentally friendly practices into cloud operations. The chapter sets the stage for exploring how cloud computing can be made more sustainable while continuing to meet the needs of users and businesses.
• 01:00 - 03:00: Challenges in Sustainable Cloud Computing The chapter discusses challenges in sustainable cloud computing, focusing on different components and taxonomy required for this paradigm or model. It continues from the previous class with a quick recap.
• 03:00 - 05:00: Components of Sustainable Cloud Computing The chapter focuses on the components of sustainable cloud computing, emphasizing the reliance of cloud service providers (CSPs) on data centers (CDCs). These data centers are critical in addressing the increasing computing demands brought about by the proliferation of IoT devices, which number in the millions and billions globally. The transcript suggests a discussion on how these data centers cater to such extensive computational needs.
• 05:00 - 08:00: Energy Management and Thermal Aware Scheduling The chapter titled 'Energy Management and Thermal Aware Scheduling' dives into the growing demand and expansion of cloud services, which has led to the increased capacity of distributed data centers. This involves the addition of more servers in these centers, consequently resulting in higher heat generation and power consumption. The chapter aims to explore strategies to manage energy and maintain thermal efficiency amidst these challenges.
• 08:00 - 11:00: Virtualization and Capacity Planning The chapter titled 'Virtualization and Capacity Planning' addresses the challenges related to the sustainability of computing systems, particularly cloud computing, due to cooling mechanisms and the physical constraints of semiconductor devices.
• 11:00 - 14:00: Renewable Energy and Waste Heat Utilization The chapter discusses the relationship between technological advances, specifically in computational power and compact technology like VLSI, and how these advances impact renewable energy and waste heat utilization. It highlights the limitations of current technologies and the challenges in further compacting or efficiently packing technology.
• 14:00 - 17:00: Taxonomy and Metrics of Sustainable Cloud Computing The chapter discusses the implications of creating compact circuits, primarily focusing on the challenge of heat generation and dissipation. The narrator highlights that heat is a critical limiting factor in developing sustainable cloud computing technology. These considerations tie into broader thermodynamic principles, although the speaker acknowledges a lack of expertise in that field. Overall, the chapter underscores the importance of managing thermal outputs to ensure sustainability in technology applications.
• 17:00 - 20:00: Application Design and Resource Management The chapter "Application Design and Resource Management" explores the architecture of data centers with a focus on cooling mechanisms. In environments such as GPUs or HPCs, extensive heat sinks are necessary to dissipate heat and maintain optimal performance. The discussion highlights the importance of effective cooling in modern data centers, especially those containing large servers or HPC servers.
• 20:00 - 23:00: Conclusion and References The chapter discusses the costs associated with IT infrastructure and environmental control mechanisms, like heating and cooling systems. It emphasizes that the expenses for these systems are comparable to the core costs of the IT infrastructure itself.

Lecture 55 Sustainable Cloud Computing - II Transcription

• 00:00 - 00:30 [Music] [Music] hello will continue our discussion on this cloud computing more specifically where last class we are discussing about sustainable cloud computing so we will continue our discussion on sustainable cloud computing right so what we ah see the concept
• 00:30 - 01:00 primarily the sustainable cloud computing thing and along with that we will look at that what are the different components taxonomy for ah needed for this sustainable cloud computing ah paradigm or sustainable cloud computing asp means model of the this cloud company so this is we are continuing from the last class just to start with a a with a quick recap what we have seen in the last class
• 01:00 - 01:30 this csps what we see that heavily rely on these cdcs or the data centers which are are there to cater this ever increasing need of the user especially in today's world with so many or a millions and billions of iot devices are in place and so this overall computing requirement is
• 01:30 - 02:00 increasing ah exponentially we may say right so to cater to this this cloud service providers are increasing their capacity like in terms of increasing number of their distributed data centers packing more servers into the data centers and which in turn lead to more heat generation more power requirement ah which in turn have more
• 02:00 - 02:30 cooling mechanisms into the place right so all those things ah brings about a ah a problem or a challenge for the sustainability of the computing system in our case cloud computing system right if you see in a more holistic manner so if you see that over years or over decades this the overall size of this ah our chips and semiconductor devices has
• 02:30 - 03:00 reduced to a great extent whereas the computational power has increased right in other terms if we if we see that ah other than there is a limit for the particular say technology like vlsi technology or whatever you look at it ah that that how much compactness or how much packing you can done ah on the other hand once you
• 03:00 - 03:30 once you make this type of ah very compact circuits is generate a tremendous amount of ah heat right which need to be dissipated right so that also in ah if you loo if we see in the thermodynamics term i am not a nowhere expert in that area but in but intuitively we see that that that is also a limiting factor how much heat you can ah tolerate right so if you see that ah
• 03:30 - 04:00 say gpus or hpc type of architecture so there is a enormous amount of ah this ah so called cooling mechanism heat sinks which which is required to cool the systems so that it can perform at its optimal level right so rather in modern day data centers specially housing large servers or hpc type of servers
• 04:00 - 04:30 what we see that the somewhere the cost of the this it infrastructure and cost of this environment like heating mechanism power sorry cooling mechanism power and say environmental other control things is basically leading to the same cost like if it is cost x this if infrastructure this is
• 04:30 - 05:00 somewhere it is if you look at it it is another x type of things right so that means ah in order to make this sustained sustainable or in order to this cloud computing sustainable what we want to look at we need to look at the all the factors right that starting from this ah computing wise how much how what is the how much we can go on increasing along with that how to handle this ah say cooling mechanism what to do
• 05:00 - 05:30 with this ah what we say this excess heat or sometimes what what is referred to as waste heat right what to do with this ah waste heat thing and there are definitely ah things like power and types like that right unless we take whole this whole mechanisms like computing plus if we say this ah per say this overall environment to handle the whole system into together ah
• 05:30 - 06:00 the making it sustainable is is always a challenging job so what we see that financial and carbon footprint related to the cost of running of such infrastructure negatively impact the sustainability right so at some point it will basically say that this is the ah you need you cannot cross this things right ah like if i have a ah say ten co train cos ten feet cross ten feet a room or maybe ah a
• 06:00 - 06:30 particular psi x x cross y ah a room with ac of so much ac and things like that so the number of such server even the space allowed is something related with the how much heat generated and cooling is there it is not only not only that how much your space or power allows also the ah how much heat it is generating and how whether there is a way to cool it uh cool it down and type of things that is one aspects if i say that the cooling mechanism things are there then means
• 06:30 - 07:00 actually all these factors this cooling mechanisms ah heat generations ah power computing they need to be looked into a more ah integrated fashion right what this sustainable cloud computing or sustainable computing in general tries to address this so the focus is to on minimizing the energy consumption and carbon footprint and ensuring reliability of the ah these cloud data centers so this is ah the goal or one of
• 07:00 - 07:30 the major objectives of this sustainable cloud computing now ah if we look at this picture also last class we have seen so this is a typical ah model ah the conceptual model proposed by ah this in this particular ah reference in this particular journal or paper so what we see one side this ah
• 07:30 - 08:00 cloud architecture is there where we have iis pass and sas right on other things what we have other models other modules right one is that ah this power manager which takes care of this power and once we talk about the power it can be from our conventional sources right which takes which basically based on this fossil fuel etcetera or we can have renewable sources right like ah it may be solar it can be wind things like that
• 08:00 - 08:30 right so that is the things then once we have this type of ah stuff then there are lot of things that when to switch to conventional to renewable and things like that and and there are pros con in running the iphone first infrastructure at the in the whole stuff like initially we may start with the conventional sources then switch it to that renewable sources and if required come back based on the load and type of
• 08:30 - 09:00 things right so that is another thing is the cool cooling manager like temperature control heat exchanger and type of things which takes care of that so called air containing conditioning infrastructure and there are remote cdc's right where that data centers there are remote data centers where we may require to ah given all those things constant power and cooling mechanisms
• 09:00 - 09:30 other things constant so i can i can have with the load increasing or the user load increase i can basically migrate the vm from one to another and overall ah say power requirement or heat dissipation i can ah we can manage right whether that is a possible right that type of thing so if if there are remote series which to market and my migrate what should be my migration plan and things like that will come into play so all those things should hand work hand
• 09:30 - 10:00 in hand and if you look at that past thing on the left side of the things there are there are definitely other press things other than there are things for remote cdc manager ah vm resource manager green resource manager thermal profiling cooling manager these are are different type of ah applications which are running at the infrastructure itself to have a more holistic approach
• 10:00 - 10:30 now this also we started at the end we basically shown this so if you look at that overall what are the different components or what are the different aspects of this cd sdc ss scc rather ah substance sustainable cloud computing so one will is the application model right what sort of application how we have tuned the application the same application may be more energy friendly if you if you have a
• 10:30 - 11:00 more optimized way of planning it there are things for virtualizations that because of virtualization is one of the major aspects of cloud computing so how you handle that whether vm migration storage migration things like that there are things for waste heat utilization there are waste heat because which are heat generated either you throw it out whether there is a possibility of using this heat right ah like there are ah concept of vapor
• 11:00 - 11:30 evaporations right once you evaporate then basically it ah cools down and things like that whether it can be used for this type of things thermal aware scheduling right so scheduling mechanism architecture ah which are thermal aware right so whatever ah architecture the scheduling whatever we are following in our mechanism whether these are thermal aware like the one is that doing efficiency in
• 11:30 - 12:00 different terms and other way looking at it that whether like it can be thermal aware there can be renewable energy which plays a big role um like energies source ah storage device which location etcetera and these are the renewable energy ah there can be renewable energy and once renewable energy is there so we require some sort of energy management not only renewable energy any energy stuff we require a energy management
• 12:00 - 12:30 right so energy management to the in first it infrastructure itself like processor processor memory storage or the cooling mechanism or network or even when to switch from renewable to the our standard [Music] energy sources and things like that right whether if the load is increased whether it is good to have ah work on with ah our standard standard
• 12:30 - 13:00 energy sources which are which are [Music] which get the energy from this fossil fuel etcetera which may be which may be more ah quickly to deliver power or renewable energy that that that those things are to be worked out and of course make the make this whole so run properly we need to do a capacity planning right that is required for any any things because now if you see this cloud computing it is not only putting some infrastructure and static set it
• 13:00 - 13:30 runs so it is much more than that why we are doing that so that we can make this sort of infrastructure sustainable right so capacity planning so power infrastructure i t sources ah work loads based on that we need to plan the capacity things right not only that we need to have predictive models like what we are heading to ah if this is my requirement down the line if we going to ah next
• 13:30 - 14:00 stage of things whether whether there is requirement increases and what sort of requirements are there and things like that so that is that is ah important for capacity planning and doing for looking for the in something which is something futuristic ah man so as we discussed so application model ah the application
• 14:00 - 14:30 in this the application model plays a vital role and the efficient structure of the application can improve the energy efficiency of the cloud data centers right application model can be data parallel function parallel ah message parallel the what way we are trying to look at whether it may be data parallel it can be ah function parallel that means parallelism in executing the function or it can be message passing and things like that so based on type of application we can have different application model and this
• 14:30 - 15:00 model can be ah what we say energy aware or thermal aware ah type of things right how much consumption of the energy how much heat generates and things like that another aspect is the energy management as we are talking about so energy consumption of the processor memory storage ah cooling of the things where some from some reported work as the document we are following have shown the ah reference in the previous slides
• 15:00 - 15:30 so what some study shows that ah for processor it is 45 percent next comes the cooling ah a 20 percent ah and this memory storage networking these are the 15 10 10 type of things right so we need to look into this when we look for the energy management who is basically ah main tickers or the consumer of the things appropriately we need to look into the things right 10 percent cooling
• 15:30 - 16:00 requirement ah if we can reduce it may ah reduce more than the same ten percent if we want to reduce for the storage etcetera right so power regulation approaches increase the energy consumption during work load execution which affects the resource utilization of the series series right so if we have a power regulations ah across the thing without so when at the peak time ah the overall performance may be affected right so
• 16:00 - 16:30 there are things like ah we have ah last case also we discussed that dvfs attempts to solve the problem of resource utilization but switching of resources between high scaling and low scaling modes increases the response time and the ah service delivery delay right so too many switching may have effect on the sla per se right so again putting server in sleep mode then wake up things when demands come
• 16:30 - 17:00 also is challenging thus improving energy efficiency in looking at different aspects may have a better way of sustainability and keeping the resource utilization reliability performance of the server in a [Music] within the threshold or in a optimal way so thermal aware scheduling as we discussed so component of thermal seedling are the architecture and
• 17:00 - 17:30 scheduling mechanisms right so can be different architecture single core multi core so heating problem during the execution of the workload reduces the efficiency of the cloud data centers right to solve this heating problem of the cdc thermal scheduling is design right so ah so what we see that it ah present day thermal average scheduling or whatever you use is mostly focused on reducing power uses efficiency right and
• 17:30 - 18:00 but p u e reaction may not ah reduce the total cost of ownership always anyway that is another ah [Music] stuff to look at but nevertheless thermal aware scheduling is again in need and becoming popular for sustaining this cloud infrastructure virtualization because the whole cloud is ah primarily concept is based on virtualization so during execution vm migrates and perform to balance the load
• 18:00 - 18:30 etcetera right due to lack of onsite renewable energy work load can be transferred here so this overall ah now once you work transfer the vm from one to another so there are lot of ah implication to that we have seen that how v m migration can be looked into so though this is also one major ah component to in to be looked into in realization of the sustainable cloud computing capacity planning as we discussed this
• 18:30 - 19:00 cloud service provider must innovate ah must involve an efficient and organized capacity planning mechanism right ah to attain expected return on investment because the if you see the csps point of view so they want to look at the whole thing as a investment right one one major component is the investment and there is a return on investment there are definitely sla violations to be avoided customer satisfaction and customer what we say ah likiness for the
• 19:00 - 19:30 infrastructure or the for the csp is all those factors but nevertheless there should be a return on investment right that whatever is there and investment in most of the cases are ah has to be done ah up front and the utilization slowly grows means grows up right or slowly the growth utilization slowly increases and things like that so they have to need appropriate planning or appropriate
• 19:30 - 20:00 capacity um planning not in terms of hard resources also in terms of that ah what we say regular resources right like power how to handle this waste energy or waste heat what should be the cooling mechanism so it should be more ah adaptive and things like that and we need there is a need to do a ah full flex capacity planning so there is
• 20:00 - 20:30 a need of effective capacity planning for data storage and the processing also right so it is not only the computing things like if we when we talk about i t infrastructure we want to include storage networks and things like that renewable energy as we know that either solar wind and things like that so energy storage device and location and location like whether it is offsite on site how much with ah and those things becomes
• 20:30 - 21:00 important so carbon ah uses efficiency cui can be reduced by adding more renewable energy but there are a flip side that renewable energy ah getting when there is a huge demand whether by this renewable energy source can handle this demand so we need to be looked into that how to switch between the renewable energy and our standard energy sources and things like that right so cloud cdcs are required to place near the renewal resources to make cost
• 21:00 - 21:30 effectiveness and if you see different relevance energy sources it requires a large span of area whether you look at say solar ah power thing or solar cell so it is it requires a sufficient speed or even wind ah energy where you require need to place those all those wind mill in a different places and waste is utilization ah the cooling mechanism of the heat transfer model plays an important role to utilize the
• 21:30 - 22:00 wastage things right so what is happening that lot of heats are being generated by the infrastructure now this heat either this ah what is in in our standard practice the cooling mechanisms basically support this ah heating thing right basically ah not i should not say support basically take care of this heating mechanism and cools it down whether there is a way to utilize this waste heat right so this
• 22:00 - 22:30 vapor absorption based free cooling techniques can help in reducing the cooling expenses right so there are different techniques which are used so energy efficiency can be improved by reducing the energy using the ah cooling so it can be if the waste heat is utilized in some way or other then it may help in ah better cooling mechanism so now we with this basic model in concept we try to see there in a another ah
• 22:30 - 23:00 the taxonomy the overall taxonomy of the sustainable cloud computing and with respect to the mostly with respect to the this different aspects and components or the features what we just discussed right so needless to say that which huge purification of iot devices and means extreme uses of ah or this our it infrastructure for different
• 23:00 - 23:30 other purpose like several location based service or mapping services ah video conferencing online feeding of live streaming of different events and games etcetera so it became a huge need of computation and there is a with cloud computing in place there is a there is a obvious fallback to this cloud computing infrastructure
• 23:30 - 24:00 right so in order to make it sustainable right so this ah cloud computing must be energy efficient and sustainable to fulfill the ever increasing user needs right so there is a ever increasing needs of the user in this cloud computing should be ah sustainable so research initiative on sustainable cloud computing can be ah categorized as follows one is that application design sustainability metric what are the different sustainability metrics are there starting from its ah
• 24:00 - 24:30 something which is performance centric to cooling to power to and different aspects of the things right ah capacity planning as we are discussing ah energy management virtualization thermal aware scheduling cooling mechanism renewable energy and waste heat ah energy ah utilization right these are all important because
• 24:30 - 25:00 most of this hm csps and data centers are working around the clock right so it is it never stops and the though the user need may be the demand may vary but the system need to adapt with the things right there are there is a need of predictive model that what is what we are going to get in future and plan accordingly so next couple of slide we will look into the different taxonomy
• 25:00 - 25:30 again presented in that reference which is given below right so those are taken from there ah i am referring that and not only that i rather encourage all of you to have a read on this particular paper or rather whatever in the references also we mentioned i mentioned so if you look at the application design so there are ah distinct component one is that looking at that ah
• 25:30 - 26:00 keyos mechanisms ah that overall application model work ah load type architecture and there are other ah components like like if you queries time cost respect ah workload management like ah what are the different critical in ah like critical interactive mechanisms or ah there are batch style or what there are different type of mechanisms right so
• 26:00 - 26:30 what what we try to look at the design of an application plays a vital role and effective structure of a of there is a small typo of an application can improve the energy efficiency right and different type of application may need different type of what we say designing right ah say
• 26:30 - 27:00 something a parallel [Music] computing type of application may be different when we have something called a streaming application and things like that right so the resource manager and the scheduler follow different approaches for application modeling so the resource manager and the scheduler follow different approaches for application modeling to make infrastructure sustainable and environmentally eco-friendly there is a need for green ict based innovative applications right
• 27:00 - 27:30 so we need to look at more applications or designs which can cater to this means energy efficient or thermal aware type of things next one is the ah what we want to discuss is the sustainability metric there is a big big big one right so one part if you see that more of the performance related metric which can be related to the cloud provider right
• 27:30 - 28:00 throughput application security and so and so forth other is the cloud user user satisfaction response time correct ah correctness reliability etcetera what the user is looking for right so this is the performance and other things ah other things like we have at the bottom ah mostly that heat waste heat utilization ah cooling mechanism ah right ah thermal availability metric energy
• 28:00 - 28:30 management matrix and these are the things right whereas ah there are issues of virtualization metric like which looks for that ah different virtualization techniques and things what should be the say ah if we have ah any any virtualization any migration issues or other things how need to be handled like that
• 28:30 - 29:00 there is a thing for capacity planning like we already discussed like ah looking at the capacity and also predicting that say uses of the infrastructure in terms of memory in terms of storage in terms of ah cpu gpus and things like that and there are renewable energy matrices for bringing down the carbon footprint and so on and so forth so these are the different ah sustainable metric which you need to
• 29:00 - 29:30 keep in mind right so once we employ or deploy some mechanisms which are which supports this sustainable cloud computing you look to need to look at that that that how effectiveness of that mechanisms by using looking at the different matrix right one or more of the matrix which are there capacity planning as we discussed is important things like with respect to
• 29:30 - 30:00 what are the different components ah like i t cooling power infrastructure etcetera with respect to i t work load application model ah things like auto scaling utility different utility functions those are things which are ah required for this capacity planning right then we have energy management things like as we discussed it can be ah overall
• 30:00 - 30:30 management of the energy so any sustain is an important factor because the energy need to be as at a optimal level so that the performance are not disturbed ah and slas are not violated or and use the minimum energy right so imp improving energy use reduces electricity bills and operational cost essential requirement of social computing are optimal software system design optimized air ventilation
• 30:30 - 31:00 and installing temperature monitoring tools for adequate resource utilization which overall which may include the energy efficiency so if you see there are things called static way of managing static management which mostly of the system level and the cpu level and there are lot of dynamic things like resource management resource consolidation configuration components which looked into the um different aspects of energy management but nevertheless is one of the major factor virtualizations
• 31:00 - 31:30 we have already discussed is an important part for the for not only for the sustainable cloud computing or cdcs but it is a it is the core of the philosophy of cloud computing right ah so this once we have this virtual machines if there is a virtual migration vm migration we have elasticity vm load balancing vm consolidation fault tolerant and scheduling all those
• 31:30 - 32:00 aspects need to be can be made [Music] in a efficient way so that the overall energy requirement or can be ah minimized ah thermal heating can be minimized and so and so forth if including waste heat type of things right thermal aware scheduling as as mentioned earlier so
• 32:00 - 32:30 consists of a the if you look at a cd a cdc or any any data any for that matter any [Music] data centers or any rack so it has a chassis racks to replace the servers to process the i t workload right to maintain temperature of the data centers right ah server produce heat during the execution of the itr cloud the processor is an important component of the server
• 32:30 - 33:00 and consume most electricity right processor things but both cooling and what we have seen ah few slide before the cooling mechanisms or other type of computing mechanisms also take or other i t infrastructure also have a huge amount of this power generation sorry ah power requirement and thermal heat generations are there so there are
• 33:00 - 33:30 different aspects like architecture heat modelling thermometer scheduling monitoring ionis and simulations and things like that then we have the cooling management so the increasing demand for computation networking storage expands the complexity right size and the energy density of series is exponentially which consumes a large amount of energy and reduces the huge amount of ah heat right so we have appropriate
• 33:30 - 34:00 cooling mechanisms which need to be in place in order that the everything works faithfully to make cdc is more energy efficient and sustainable we need an efficient cooling management systems which maintain the temperature of cdc right and then we have the redu renewable energy right different sources like it can be solar wind or in sometimes
• 34:00 - 34:30 like hydro power water generated energy which has a near zero carbon dioxide emission things like that um that can be employed and of course if we need to these then we need to have appropriate work lock schedule schedule there so what's which sort of workload to be scheduled ah what are the different sources on energy whether it should be location aware whether we can have this storage device
• 34:30 - 35:00 how to maintain with this renewable energy and things like that then finally we are having this waste heat utilization right so reuse of waste heat is becoming a solution for fulfilling energy demand in the energy conservation system so the waste heat means heat generated by the system itself which are cooled down but if we have a mechanism of utilizing this reusing this waste heat um then we in efficient we
• 35:00 - 35:30 we can basically overall ah what we say ah like overall this green computing aspect we may ah pull we may put more ah emphasis on this right so that one way of what is seen in the literature people are trying for vapour absorption based cooling system as we are discussing in the beginning of the thing can use waste heat and remove the heat while evaporating right so vapor absorption
• 35:30 - 36:00 based free cooling mechanisms can ah make the value of p u e ideal ah by neutralizing the cooling expenses so it it will have a not only monetary wise but also you can have utilization of the things right so if we look at the overall sustainable cloud computing so this ever increasing demand with ah iots and
• 36:00 - 36:30 different type of sense basic finance sensors even this huge proficient of this drone level drone operations and capturing different images and other things so we get a huge volume of data right apart from there there are streaming data and things like that so ah so we need to have more efficient so which what which dictates this ah some sort of a ah ever increasing form of cloud cloud
• 36:30 - 37:00 infrastructure which is may not be feasible up to a limit so what we need to do we need to have this all this component together and make a sustainable model right so that it sustain over time span so this what we see this next generation or the today's cloud computing ah must be energy efficient and sustainable to fulfill this sustainable component cdcs are powered by renewable energy sources kept more towards it means
• 37:00 - 37:30 this near the power generation things may be more in a cool more cool environment right there are other issues of ah environmental issues right so that that overall carbon footprint is ah reduced right so sustainability with high performance and reliability is one of the primary goal of of this cloud computing or today's cloud computing infrastructure so with this let us disc end our discussion today there are very nice references ah some of them i
• 37:30 - 38:00 have put here so i encourage you to go to these references that will be most most of the things we basically taken from these different references but i think that reading the differences will be good for you thank you