Cloud computing is a cost-effective and scalable infrastructure for running High Performance Computing (HPC) and web applications. However, the growing demand for cloud infrastructure increases the energy consumption of data centers.
Clouds are virtualized data centers offered as services on a subscription basis. Data centers consume huge amounts of energy and have high operational costs. The data centers introduce high carbon footprints due to the extensive amounts of electricity needed to cool and power multiple servers.
The power consumption of data centers has become a major issue due to energy shortages and global climate change. To save energy and reduce operational costs, cloud computing vendors should invest in green technologies.
The Need for Green Cloud Computing
Modern data centers require massive amounts of electricity to power and cool multiple internal servers. A typical datacenter with 1000 racks needs 10 Megawatt of power to operate. As a result the operational cost and the carbon footprint of data centers increases.
Large servers and disks are needed to process a growing number of high performance computing applications and data. The process has to be fast and efficient. Lowering the energy usage of data centers is a challenging and complex issue.
Cloud providers have to use green and sustainable technology. Otherwise, cloud computing can cause massive growth in energy usage. To address this problem, cloud providers need to manage data centers in an energy-efficient manner. They need to allocate cloud resources not only to satisfy Service Level Agreement (SLA) requirements, but also to reduce energy usage.
6 Cloud Features Enabling Green Computing
Technological improvements like virtualization and innovative cooling systems improve server tolerance to humidity and heat. Trends like server consolidation reduce energy consumption by around two-thirds. In addition, the technology sector uses a higher proportion of renewable energy than any other industry.
The following list reviews the main cloud computing features that enable the reduction of data center carbon emissions.
1. Dynamic provisioning
Dynamic provisioning is an effective way to deploy virtual machine instances from a centralized console or application. Cloud providers are responsible for migrating these virtual machine instances to another host in case user application requires more resources. The reasons for such deployments can be unstable application loads or the need for high service availability.
Cloud providers predict and monitor the demand for application resources and allocate them accordingly. You can consolidate applications that require fewer resources on the same server. Thus, data centers can consistently maintain active servers according to current demand. As a result, energy consumption decreases compared to the traditional provisioning approach.
2. Multi-tenant architecture
Multi-tenancy allows multiple instances of virtual machines to run on a single server. Cloud infrastructure can reduce the overall usage of energy and associated carbon emissions using multi-tenant architecture.
Software as a Service (SaaS) providers use the same infrastructure and software to serve multiple companies. This strategy is more energy-efficient than numerous copies of software installed on different infrastructure. Moreover, businesses often have different demand patterns. By controlling the demand, multi-tenancy reduces the need for extra infrastructure.
3. Server utilization
Server utilization is a measurement of the used server infrastructure during a client request or other services running on the server. Utilization depends on multiple factors like the number of requests, transmission bandwidth, the type and size of the content, and the server architecture.
Typically, on-premise infrastructure runs with very low utilization. Sometimes it can go down to 5 or 10 percent of average utilization. On the other hand, virtualization enables you to host and execute multiple applications on the same server. Thus, it drastically reduces the number of active servers, which leads to approximately 70% of utilization.
Servers running at higher utilization can process more workload with similar power usage. As a result, cloud computing can benefit from high utilization even though it results in more power consumption.
4. Data center efficiency
Cloud providers can significantly improve the Power Usage Effectiveness (PUE) of data centers by using energy-efficient technologies. This includes container-based server design, power supply optimization, and advanced water or air based cooling.
By using these technologies cloud providers can significantly improve PUE and obtain 40% more power efficiency than traditional data centers. In addition, by using virtualized services and high-speed networks, cloud computing enables companies to move their services to data centers with better PUE values.
Many companies are moving to cloud-based SaaS applications to minimize their IT costs. As cloud consumers, they must also address the issue of energy efficiency in the application level. However, the application layer receives very little attention since most applications are an upgraded version of previously developed tools.
SaaS providers have to develop software on energy-efficient infrastructure to achieve energy efficiency at the application level. There is always a trade-off between energy and performance due to the execution of software on multiple platforms and hardware. In addition, software developers must leverage different energy-efficient programming techniques during the design of their future applications.
6. Network infrastructure
Data centers can achieve energy efficiency at the network level either at the network nodes or in switches and routers. Network energy efficiency problems are often referred to as “green networking”. These issues relate to raising energy-awareness in design, devices and in network protocols.
Possible solutions include virtualization, resource consolidation, proportional computing and selective connectedness. Selective device connectedness enables a single piece of equipment to go idle for some time without affecting other network devices. Resource consolidation helps reduce global consumption by regrouping the under-utilized servers.
Cloud computing resources produce high emissions of CO2. However, cloud providers are trying their best to improve efficiency in every aspect, including infrastructure, software, and even in terms of business models.
Renewable technology adoption delivers economy-wide improvements in carbon productivity. However we cannot stay still, efficiency is not improved automatically. Organizations must carefully deploy new technologies and keep an eye on business processes and trends. Also, cloud providers need to better understand energy flows, and make consumption more transparent, especially for consumers.