Sreekumar C Pillai| Informationweek
The cloud lifecycle can be divided into four stages, viz., designing, setup, using and disposal. Green computing in the cloud context can be achieved by ensuring minimum or no impact on environment during each of these cloud lifecycle stages.
The objective is to reduce energy consumption and improve resource performance and efficiency. Green computing in a Cloud can be achieved by initiating a series of transformative actions such as improving Cloud data center design, increasing resource longevity, consolidating resources, and optimizing algorithms. Let us look at each of these in detail.
CLOUD DATA CENTER DESIGN
Data center design plays a very important role and is essential for creating an energy efficient data center with energy saving configuration.
Creating a good data center design requires attention to be paid to the following:
• Data center location:
Decide data center location factoring the data center purpose as well as availability of resources for running the data center such as cheaper electricity and skilled resources.
• Construction of the data center building: While designing the data center, electrical systems must consider the data center landscape and wherever possible leverage the natural lighting. Other factors to be considered are local availability of renewable energy, using outside air for cooling or locating the systems where the heat they produce may be used for other purposes.
• Resources: The selection of a resource to be used in the data center must be based on lower energy consumption by the resource. This will help reduce the running costs. Other steps include using efficient air management, cooling and electrical systems while designing data centers. Terminal servers have also been used in green computing. Users connect from their terminal to a central server. Though all the computing is done at the central server, the user experiences the operating system on the terminal. Combining of terminal servers with thin clients that use about one-eighth of the energy of a normal workstation, result in decrease of both energy costs and consumption.
• Configuration: Align all IT processes and systems with the core principle of sustainability. The IT systems should be designed from the cloud perspective and must help leverage the benefits of the cloud. Many operating systems provide Advanced Configuration and Power Interface (ACPI), an open industry standard that allows an operating system to directly control the power saving aspects of its underlying hardware.
Data center must be designed so as to ensure appropriate cooling throughout the data center. It is crucial to use energy efficient cooling systems in the data center and intelligent systems for temperature control within the data center. One must also provide for alternative power supply for use in case of disaster, setup processes for regular system maintenance and conduct periodic checks to ensure that the systems are functioning properly. The above considerations will help arrive at a good design resulting in better space utilization and increased performance and efficiency.
INCREASING RESOURCE LONGEVITY
Each resource has a lifespan. One can contribute to green computing by increasing the resource longevity by including upgradability and modularity. For example, it is said that manufacturing a new PC makes a bigger ecological footprint than making a new RAM module to upgrade an existing one.
RESOURCE CONSOLIDATION
Another important factor that plays a major role in creating an energy efficient data center is resource consolidation. This can be achieved by virtualization, and consolidation of other resources that cannot be virtualized. This gives flexibility of sharing the available resources and allocating then depending on the business need. Resource consolidation by virtualization helps to reduce the quantum of energy consumed. Other ways by which resource optimization can be achieved are shutting down resources when not needed, reuse and recycling.
OPTIMIZATION
The efficiency of an algorithm has an impact on the resources required for a computing function and at times one may need to do trade-offs while writing programs. Use efficient algorithms. For example, use fast search algorithms such as hashed or indexed search algorithms instead of slow linear search algorithms. Cost optimization can be achieved by using algorithms to route the data to data center where electricity is cheaper. In case data center is facing warm weather, traffic could be routed away to cut energy usage, allowing the servers to shut down and avoid using the air conditioning.
We have seen above how different measures such as cloud data center design, increasing resource longevity, resource consolidation and optimization help achieve green computing in the cloud context.
There is scope for further reducing the carbon footprint in the Cloud. Research is being done in areas such as optimization of data center hardware and software, improving power supply chain and data center cooling technologies.
