Salesforce’s chief executive officer has urged coders to include climate change in their software.
According to Forbes, “Even seemingly unrelated activities, such as hosting or building a website, and designing software can have significant climate impacts,” wrote Salesforce EVP and chief impact officer Suzanne DiBianca last week.
She explained that the focus so far has been on reducing energy consumption in data centers Moving electrical grids off fossil fuels. “Now,” she added, “coders, designers, and AI experts are ready for the same push in software, cryptoproof of work, or AI compute power.”
Few coders know how to make their creations more environmentally friendly, even though many want to, said Ms. Schwartz.
DiBianca quoted a Salesforce survey that included more than 1,000 technologists from the United States (UX designers), United Kingdom (Software Developers) and Australia (IT Operations Managers). The survey revealed 75% of UX developers, software managers, and IT operations managers wanted software to be less damaging to the environment. Yet, half of them don’t have a clue about how to minimize environmental damage.
She said that more than one-third (34%) admitted to not considering carbon emissions “rarely or ever” when they type a new line in code.
She stated that convincing the management to commit would be one of her biggest obstacles in creating a major shift in how software is built. According to the Salesforce study, 76% don’t think sustainable software is a must-have.
DiBianca wrote: “Without commitment from the top”, “technologists must adapt in a vacuum,” without the training, skills, or mandate needed to make a real difference.
Lagging in Sustainability
Abhijit Snil, senior analyst at Unisys, noted that sustainability in IT is not as developed as it is for greening software. Forrester Research, a company that conducts national market research with headquarters in Cambridge Mass.
He told TechNewsWorld that many organizations have struggled to implement sustainability measures for application development and software.
He said that “a lot of metrics are still in development and difficult to measure.” “For instance, how can we measure the carbon impact of large enterprise apps?”
Sunil added, “We are beginning to see industry groupings develop guidelines and best practices.” “One of the examples is the Green Software Foundation“, which has been publishing guidelines and best practices including ways to measure software sustainability.”
Forrester’s clients have cited the Green Software Foundation in their search for sustainable software.
Another organization is SustainableIT.org This year, the company released its first ever set of environmental standards for measuring the impact of IT operations, including the building, management, and running of IT systems.
These standards define and provide metrics for energy consumption, waste and emissions. They are intended to help organizations accurately and consistently gauge the environmental impact of IT, set goals, and track progress in improving sustainability.
AI: Greener Code
DiBianca said that artificial intelligence is a great way to create greener code.
A recent report from Salesforce found that AI’s power is not being fully utilized when it comes green code. Sixty-percent of leaders don’t use AI or automation to make software development more energy efficient.
Crispin Cowan, former professor of computer science, startup founder, and current staff engineer of AT&T, said that AI is among the least green forms of computation. TaniumKirkland, Washington, is the home of a maker endpoint management and a security platform.
“It’s vastly computation-intensive,” said he to TechNewsWorld. AI-driven research could lead to new energy-saving technologies in the future. But for now, it will cause IT’s electrical consumption to increase.
Mike Parkin added that AI is likely to have a greater impact on the design of energy-efficient infrastructure rather than on code efficiency. Vulcan CyberMaker of a Cyber Risk Management Platform, Tel Aviv, Israel
He told TechNewsWorld that there are many ways to improve code efficiency, but they must be done while keeping in mind important factors such as security and usability.
Optimize Software with Emissions in mind
DiBianca recommended that software be designed with emissions in consideration.
She wrote, “When creating a new product, or website, designers have the ability to create better, quicker, and more sustainably user experiences.” “Designers can select the most sustainable option by default, making sustainability easy and desirable for the users.”
She explained, “Even small adjustments to the image size, colors, and type options have large effects.”
Parkin said that code optimization was an art. It involves balancing factors such as efficiency, speed, memory footprint and others to achieve design goals.
He said that while code used to be highly optimized in the past, as computing resources improved, they have started to focus less on high efficiency.
He continued, “We can work with gigabytes and high-level abstraction languages.” “It is unlikely we will see a resurgence of assembler programming to squeeze maximum performance out of limited resource.”
Cowan said that efficient code will help to save energy. “Turn the optimization level up and cut fancy graphics, such as animations, shadows, and other effects down. [and] He recommended that polling be done as rarely as possible.
“Efficient code reduces the amount of energy consumed by IT,” said he. This efficiency gain, however, can be multiplied a thousandfold by using these CPU cycles to improve business process.
He said that using a few milliwatts in order to calculate how much a truck could save on distance, time and fuel by changing its route was a big win for reducing greenhouse gasses and carbon footprint.
Hardware Infrastructure: Green Benefits
DiBianca said that equipping technologists and leaders with the correct tools will not only help them achieve meaningful cost savings and efficiencies, but they can also reduce emissions at the pace and scale required by the planet.
Parkin, however, argues that coding for an eco-friendly environment is fascinating but it does not include a key component that it requires to be successful: instrumentation.
The coders cannot make the code more energy-efficient if they do not have a clear understanding of how it performs from an energy perspective.
“It is likely that there are more gains to be had in improving the hardware infrastructure upon which their code relies than by optimizing the individual function calls in order to lower the power load,” said he.
He concluded that “a few percent gained, for example, in improving cooling efficiency in the Data Center is almost certain to be less expensive and more efficient than spending hours optimising code for a 0.5 percent gain per transaction,”
