Modern
IT systems rely on a complicated mix of people, networks, and hardware; as such, a green computing initiative ideally covers these areas. A solution may also need to address end user satisfaction, management restructuring, regulatory compliance, and return on investment (ROI). There are also fiscal motivations for companies to take control of their own power consumption; "of the power management tools available, one of the most powerful may still be simple, plain, common sense."
Product longevity Gartner maintains that the PC manufacturing process accounts for 70% of the natural resources used in the life cycle of a PC. In 2011, Fujitsu released a
life-cycle assessment (LCA) of a desktop that show that manufacturing and end of life accounts for the majority of this desktop's ecological footprint. Therefore, the biggest contribution to green computing usually is to prolong the equipment's lifetime. A recent life-cycle assessment comparing a desktop and a laptop for a four-year use case with similar performance found total carbon footprints of 679.1 kg CO2e for the desktop versus 286.1 kg CO2e for the laptop; for both systems, manufacturing was the largest contributor, followed by the use phase. Another report from Gartner recommends to "Look for product longevity, including upgradability and modularity." For instance, manufacturing a new PC makes a far bigger
ecological footprint than manufacturing a new
RAM module to upgrade an existing one.
Data center design Data center facilities are heavy consumers of energy, accounting for between 1.1% and 1.5% of the world's total energy use in 2010. Energy efficient data center design should address all of the energy use aspects included in a data center: from the IT equipment to the HVAC (Heating, ventilation and air conditioning) equipment to the actual location, configuration and construction of the building. The U.S. Department of Energy specifies five primary areas on which to focus energy efficient data center design best practices: • Information technology (IT) systems • Environmental conditions • Air management • Cooling systems • Electrical systems Additional energy efficient design opportunities specified by the U.S. Department of Energy include on-site electrical generation and recycling of waste heat. Energy efficient data center design should help to better use a data center's space, and increase performance and efficiency.
Software and deployment optimization Algorithmic efficiency The efficiency of algorithms affects the amount of computer resources required for any given computing function and there are many efficiency trade-offs in writing programs. Algorithm changes, such as switching from a slow (e.g. linear)
search algorithm to a fast (e.g. hashed or indexed) search algorithm can reduce resource usage for a given task from substantial to close to zero. In 2009, a study by a physicist at
Harvard estimated that the average Google search released 7 grams of carbon dioxide (CO2). However, Google disputed this figure, arguing that a typical search produced only 0.2 grams of CO2.
Resource allocation Algorithms can also be used to route data to data centers where electricity is less expensive. Researchers from MIT, Carnegie Mellon University, and Akamai have tested an energy allocation algorithm that routes traffic to the location with the lowest energy costs. The researchers project up to 40 percent savings on energy costs if their proposed algorithm were to be deployed. However, this approach does not actually reduce the amount of energy being used; it reduces only the cost to the company using it. Nonetheless, a similar strategy could be used to direct traffic to rely on energy that is produced in a more environmentally friendly or efficient way. A similar approach has also been used to cut energy usage by routing traffic away from data centers experiencing warm weather; this allows computers to be shut down to avoid using air conditioning. Larger server centers are sometimes located where energy and land are inexpensive and readily available. Local availability of renewable energy, climate that allows outside air to be used for cooling, or locating them where the heat they produce may be used for other purposes could be factors in green siting decisions. Approaches to actually reduce the energy consumption of network devices by proper network/device management techniques have been surveyed Bianzino, et al. The authors grouped the approaches into 4 main strategies, namely (i) Adaptive Link Rate (ALR), (ii) Interface Proxying, (iii) Energy Aware Infrastructure, and (iv) Maximum Energy Aware Applications.
Virtualizing Computer virtualization refers to the abstraction of computer resources, such as the process of running two or more logical computer systems on one set of physical hardware. The concept originated with the IBM
mainframe operating systems of the 1960s, and was commercialized for
x86-compatible computers, and other computer systems, in the 1990s. With virtualization, a system administrator can combine several formerly physical systems as virtual machines on one powerful system, thereby conserving resources by removing need for some of the original hardware and reducing power and cooling consumption. Virtualization can assist in distributing work so that servers are either busy or put in a low-power sleep state. Several commercial companies and open-source projects now offer software packages to enable a transition to virtual computing.
Intel Corporation and
AMD have also built proprietary
virtualization enhancements to the x86
instruction set into each of their CPU product lines, in order to facilitate virtual computing. New virtual technologies, such as
operating system-level virtualization can also be used to reduce energy consumption. These technologies make a more efficient use of resources, thus reducing energy consumption by design. Also, the consolidation of virtualized technologies is more efficient than the one done in
virtual machines, so more services can be deployed in the same physical machine, reducing the amount of hardware needed.
Terminal servers Terminal servers have also been used in green computing. When using the system, users at a terminal connect to a central server; all of the actual computing is done on the server, but the end user experiences the system operating as if it were on the terminal. These can be combined with
thin clients, which use up to 1/8 the amount of energy of a normal workstation, resulting in a decrease of energy costs and consumption. There has been an increase in using terminal services with thin clients to create virtual labs. Examples of terminal server software include
Terminal Services for Windows and the
Linux Terminal Server Project (LTSP) for the
Linux operating system. Software-based remote desktop clients such as
Windows Remote Desktop and
RealVNC can provide similar thin-client functions when run on low power hardware that connects to a server.
Data Compression Data compression, which involves using fewer bits to encode information, may also be used in green computing depending on the structure of the data. Since it is highly data specific, data compression strategies may result in using more energy or resources than necessary in some cases. However, choosing a well-suited compression algorithm for the dataset can yield greater power efficiency and reduce network and storage requirements. There is a tradeoff between compression ratio and energy consumption. Deciding whether or not this is worthwhile depends on the dataset's compressibility. Compression improves energy efficiency for data with a compression ratio much less than roughly 0.3, and hurts for data with higher compression ratios.
Power management The
Advanced Configuration and Power Interface (ACPI), an open industry standard, allows an operating system to directly control the power-saving aspects of its underlying hardware. This allows a system to automatically turn off components such as
monitors and
hard drives after set periods of inactivity. In addition, a system may
hibernate, when most components (including the
CPU and the system RAM) are turned off. ACPI is a successor to an earlier Intel-Microsoft standard called
Advanced Power Management, which allows a computer's
BIOS to control power management functions. Some programs allow the user to manually adjust the voltages supplied to the CPU, which reduces both the amount of heat produced and electricity consumed. This process is called
undervolting. Some CPUs can automatically undervolt the processor, depending on the workload; this technology is called "
SpeedStep" on Intel processors, "
PowerNow!"/"
Cool'n'Quiet" on AMD chips,
LongHaul on
VIA CPUs, and
LongRun with
Transmeta processors.
Data center power Data centers, which have been criticized for their extraordinarily high energy demand, are a primary focus for proponents of green computing. According to a
Greenpeace study, data centers represent 21% of the electricity consumed by the IT sector, which is about 382 billion kWh a year. Data centers can potentially improve their energy and space efficiency through techniques such as storage consolidation and virtualization. Many organizations are aiming to eliminate underused servers, resulting in lower energy usage. The U.S. federal government set a minimum 10% reduction target for data center energy usage by 2011. Advocates have argued that cryptocurrency can help to drive investment in green energy.
Operating system support Microsoft Windows has included limited
PC power management features since
Windows 95. These initially provided for stand-by (suspend-to-RAM) and a monitor low power state. Further iterations of Windows added hibernate (suspend-to-disk) and support for the
ACPI standard.
Windows 2000 was the first NT-based operating system to include power management. This required major changes to the underlying operating system architecture and a new hardware driver model. Windows 2000 also introduced
Group Policy, a technology that allowed administrators to centrally configure most Windows features. However, power management was not one of those features. This is probably because the power management settings design relied upon a connected set of per-user and per-machine binary registry values, effectively leaving it up to each user to configure their own power management settings. This approach, which is not compatible with Windows Group Policy, was repeated in
Windows XP. The reasons for this design decision by Microsoft are not known, and it has resulted in heavy criticism. Microsoft significantly improved this in Windows Vista by redesigning the power management system to allow basic configuration by Group Policy. The support offered is limited to a single per-computer policy. Windows 7 retains these limitations but includes refinements for
timer coalescing, processor power management, and display panel brightness. The most significant change in Windows 7 is in the user experience. The prominence of the default High Performance power plan has been reduced with the aim of encouraging users to save power. Third-party
PC power management software for adds features beyond those built-in to the Windows operating system. Most products offer
Active Directory integration and per-user/per-machine settings with the more advanced offering multiple power plans, scheduled power plans, anti-insomnia features and enterprise power usage reporting.
Linux systems started to provide laptop-optimized power-management in 2005, with power-management options being mainstream since 2009.
Power supply Desktop
computer power supplies are in general 70–75% efficient, dissipating the remaining energy as heat. A certification program called
80 Plus certifies PSUs that are at least 80% efficient; typically these models are drop-in replacements for older, less efficient PSUs of the same form factor. As of July 20, 2007, all new Energy Star 4.0-certified desktop PSUs must be at least 80% efficient.
Storage Smaller form factor (e.g., 2.5 inch)
hard disk drives often consume less power per gigabyte than physically larger drives. Unlike hard disk drives,
solid-state drives store data in flash memory or
DRAM. With no moving parts, power consumption may be reduced somewhat for low-capacity flash-based devices. As hard drive prices have fallen, storage farms have tended to increase in capacity to make more data available online. This includes archival and backup data that would formerly have been saved on tape or other offline storage. The increase in online storage has increased power consumption. Reducing the power consumed by large storage arrays, while still providing the benefits of online storage, is a subject of ongoing research.
Video card A fast
GPU may be the largest power consumer in a computer. Energy-efficient
display options include: • No video card – use a shared terminal, shared
thin client, or
desktop sharing software if display is required. • Use motherboard video output – typically low 3D performance and low power. • Select a GPU based on low idle power, average wattage, or
performance per watt.
Display Unlike other display technologies,
electronic paper does not use any power while displaying an image.
CRT monitors typically use more power than LCD monitors. They also contain significant amounts of lead.
LCD monitors typically use a
cold-cathode fluorescent bulb to provide light for the display. Most newer displays use an array of
light-emitting diodes (LEDs) in place of the fluorescent bulb, which further reduces the amount of electricity used by the display. Fluorescent back-lights also contain mercury, whereas LED back-lights do not. A
light-on-dark color scheme, also called
dark mode, is a
color scheme that requires less energy to display on new display technologies, such as
OLED. This positively impacts battery life and energy consumption. While an OLED will consume around 40% of the power of an LCD displaying an image that is primarily black, it can use more than three times as much power to display an image with a white background, such as a document or web site. This can lead to reduced battery life and increased energy use, unless a light-on-dark color scheme is used. A 2018 article in
Popular Science suggests that "Dark mode is easier on the eyes and battery" and displaying white on full brightness uses roughly six times as much power as pure black on a Google Pixel, which has an OLED display. Apple's
iOS 13 and
iPadOS 13 both feature a light-on dark mode, which would allow third-party developers to implement their own dark themes. Google's
Android 10 features a system-level dark mode.
Materials recycling Recycling computing equipment can keep harmful materials such as lead, mercury, and hexavalent chromium out of
landfills, and can replace equipment that otherwise would need to be manufactured, saving further energy and emissions. Computer systems that have outlived their original function can be re-purposed, or donated to various charities and non-profit organizations. However, many charities have recently imposed minimum system requirements for donated equipment. Additionally, parts from outdated systems may be salvaged and recycled through certain retail outlets and municipal or private recycling centers. Computing supplies, such as
printer cartridges,
paper, and
batteries may be recycled as well. A drawback to many of these schemes is that computers gathered through recycling drives are often shipped to
developing countries where environmental standards are less strict than in North America and Europe. The
Silicon Valley Toxics Coalition has estimated that 80% of the post-consumer e-waste collected for recycling is shipped abroad to countries such as
China and
India. In 2011, the collection rate of e-waste remained low, even in the most ecology-responsible countries like France. In the U.S., e-waste collection was at a 14% annual rate between electronic equipment sold and e-waste collected for 2006 to 2009. The recycling of old computers raises a privacy issue. The old storage devices still hold private information, such as emails, passwords, and credit card numbers, which can be recovered simply by using software available freely on the Internet. Deletion of a file does not actually remove the file from the hard drive. Before recycling a computer, users should remove the hard drive, or hard drives if there is more than one, and physically destroy it or store it somewhere safe. There are some authorized hardware recycling companies to whom the computer may be given for recycling, and they typically sign a non-disclosure agreement.
Cloud computing Cloud computing may help to address two major ICT challenges related to green computing – energy usage and
embodied carbon. Hyperscale data centers such as those operated by
AWS,
Azure, and
GCP can benefit from economies of scale, and
virtualization, dynamic provisioning environment, multi-tenancy and
green data center approaches can enable more efficient resource allocation. Organizations may be able to reduce their direct energy consumption and carbon emissions by up to 30% and 90% respectively by moving certain on-premises applications into the public cloud. However, critics point to shortcomings in the carbon tracking and management tools provided by major cloud providers. GreenOps, also known as DevGreenOps, DevSusOps or DevSustainableOps, is emerging as a framework to include sustainability into cloud management. Carbon-aware computing and grid-aware computing can form part of a GreenOps approach. This includes techniques like demand shifting, which means moving computational workloads to locations or times of day with cleaner energy in the grid. Demand shaping is a similar technique, which focuses on adjusting workloads according to the amount of clean energy currently available.
Edge computing New technologies such as
edge and
fog computing are a solution to reducing energy consumption. These technologies allow redistributing computation near its use, thus reducing energy costs in the network. Furthermore, having smaller data centers, the energy used in operations such as refrigerating and maintenance is reduced.
Remote work Remote work using
teleconference and
telepresence technologies is often implemented in green computing initiatives. The advantages include increased worker satisfaction, reduction of
greenhouse gas emissions related to travel, and increased profit margins as a result of lower overhead costs for office space, heat, lighting, etc. The average annual energy consumption for U.S. office buildings is over 23 kilowatt hours per square foot, with heat, air conditioning and lighting accounting for 70% of all energy consumed. Other related initiatives, such as
Hoteling, reduce the square footage per employee as workers reserve space only when needed. Many types of jobs, such as sales, consulting, and field service, integrate well with this technique.
Voice over IP (VoIP) reduces the telephony wiring infrastructure by sharing the existing Ethernet copper. VoIP and phone extension mobility also made
hot desking more practical. Wi-Fi consume 4 to 10 times less energy than 4G.
Telecommunication network devices energy indices In 2013 ICT energy consumption, in the US and worldwide, was estimated respectively at 9.4% and 5.3% of the total electricity produced. The energy consumption of ICTs is today significant even when compared with other industries. Some studies have tried to identify the key energy indices that allow a relevant comparison between different devices (network elements). This analysis was focused on how to optimise device and network consumption for carrier telecommunication by itself. The target was to allow an immediate perception of the relationship between the network technology and the environmental effect. These studies are at the start and further research will be necessary.
Supercomputers The
Green500 list was first announced on November 15, 2007, at SC|07. As a complement to the TOP500, the listing of the Green500 began a new era where supercomputers can be compared by performance-per-watt. As of 2019, two Japanese supercomputers topped the Green500 energy efficiency ranking with performance exceeding 16 GFLOPS/watt, and two IBM AC922 systems followed with performance exceeding 15 GFLOPS/watt. ==Education and certification==