DEV Community: Green Software Foundation

MEET A STEERING COMMITTEE MEMBER: YUSUKE KOBAYASHI OF NTT DATA

Green Software Foundation — Mon, 07 Feb 2022 23:26:32 +0000

Yusuke, the Manager Green Innovation at NTT DATA believes that we should motivate developers and users to benefit from innovative IT technologies, but without forgetting to consider their environmental impact.

*Tell us about your career path, interest in green software and your journey to GSF *
Working as a data scientist for more than ten years, I have discovered our clients' issues in various fields such as corporate, financial, and public sectors through data analysis and proposed business improvement plans. I also developed AI technology for infrastructure to solve social issues as R&D. I believe IT technology can solve social problems and make our world more convenient. It also has a responsibility in terms of carbon emissions. I want to contribute to the challenge of balancing development and sustainability from the perspective of data science while taking responsibility as an IT engineer.

*What do you, as an individual, expect to achieve by working with the GSF and in green software? *
Our familiar IT systems and services emit carbon. I want to spread this knowledge to help people understand and choose greener software. I also want to contribute to the development of tools and standards for realizing such green software from the data science point of view.

*What obstacles do you see to the cause of green software? How can we overcome them? *
New innovative IT technologies and their benefits make people blinded so they forget to consider whether they are green or not. However, because the latest technologies solve other problems at the same time, it is important not to deny such developments. But we should motivate developers and users to aim for coexistence. We can educate them to go green by practicing the basic steps of measurement and take action for reduction.

Any other matters you would like to share with us in the cause of green software?
Using software consumes a great deal of energy and resources. Software development is also responsible for a lot of carbon emissions. I want to help establish standards that correctly reflect the reduction efforts in software development and standardize them so that all developers know whether they are developing green software. I want to actively work on building tools and best practices to support this.

Check out what our other Steering Committee members say about green software
Jeff Sandquist of Microsoft

Elise Zelechowski of Thoughtworks

Erica Brescia of GitHub

Santiago Fontanarrosa of Globant

Dan Lewis-Toakley of Thoughtworks

Toru Shimogaki of NTT Data

Jose Lopez Dominguez of Globant

This article is licenced under Creative Commons (CC BY 4.0)

A DEMAND CURVE FOR COMPUTE

Green Software Foundation — Mon, 07 Feb 2022 23:02:14 +0000

Once you begin thinking of computing as an essential utility, we can consider different ways to provide a commodity like computing.

In an earlier post, Understanding energy market trends at the layer below the internet stack, we looked at the “stack below the internet”, and learned that trends in economics of renewables make the idea of a fossil free internet both desirable and achievable. In this post, we’ll show how you might take ideas like demand curves and different ways to generate a commodity like electricity, and use them to help us think about the different ways to provide a commodity like computing.

First, a recap of the last post
For those who haven’t read the last post in this series, the short version is as follows:

On an electricity grid, as the amount of demand for electricity climbs, so too does the cost per unit of electricity that companies are able to sell their electricity for.

This means that some kinds of generation need to be running almost all the time and selling all power they generate, to ever have a hope of paying off the loans it took to build them. But it also means there are others that barely need to run at all in a given year, and they can still pay for themselves, because they only run when the cost of electricity is really high.

We can look at the properties of these different kinds of generation, as creating a kind of demand curve.

What this has to with computers and cloud – a demand curve for compute
This concept of a demand curve is useful when thinking about how we compose systems from different kinds of infrastructure services too.

Over the last few years, computing workloads themselves have become more portable, as people get better at packaging them up into easily deployable pieces of software, and abstracting them from the underlying hardware. This has made them more fungible.

This first happened as virtual machine images, and then more recently as containers, and in 2021, we now have a plethora of options from micro-kernels, micro-vms, and so on.

Let’s assume as an example, we have a relatively portable computing job that we want to run somewhere. It might be a WordPress website, or it might be a machine learning job we need to run on a regular basis. Because it doesn’t matter that much what it is, let’s just call it a workload, to represent that’s something we need to run, but one where we have some degree of freedom where we run it.

Matching the job to the infrastructure
Once we have a computing job that’s relatively portable, we are then able to make decisions about how we run it, and match the properties of that workload to the choice of digital infrastructure that fits it best.

To do this, it’s helpful to think about the things we might trade-off on when deciding where we run our workloads.

Cost per unit of compute
Unsurprisingly, the cost we’re prepared to pay per unit for anything is something we’re likely to make a deliberate decision about. In our case, we might think in terms of how much compute we expect to get for the money we pay. Let us call this our cost per unit of compute. It’s likely to change depending where or how we run it.

Degree of abstraction
Another thing to think about for our workload might be the degree of abstraction we are comfortable with. It helps here if we think of compute power like a fungible metered utility, the same way we think about electricity as one.

You might run the same workload on a range of different providers, who are all able to provide the computing power, but with different properties that make them a better or worse fit for it.

Let’s look at them in turn, like we did with the different forms of power generation.

Different ways to run a computing job

Bare metal: low unit cost, less abstraction, harder to scale up and down
If we know we have a very predictable, but heavy compute workload that’s running almost constantly, we might make a choice to trade-off some of the ability to scale compute capacity up and down, as long as each unit of compute is relatively cheap – you might think of this like the current generation of nuclear power on the energy demand curve in the earlier blog post.

In our world of the internet, this might look like us using a bare metal server, where there is little to no abstraction of the workload from the underlying physical servers.

But this isn’t the only scenario that describes this set of conditions. Some providers sell virtual machines, at a lower cost per unit, if you commit to paying to use them for an extended period of time.

In both cases, you’re committing to a set amount of use, on an annual or longer time frame. This is still a very common way that computing is paid for, and if any thought is given to how well a set up like this scales, it’s how well it handles an increase in required work before you need to buy another new server. We assume the server has to be running in order for there to be capacity to meet demand.

Virtual machines: higher unit cost, more abstract, easier to scale
Another scenario might be where you have a workload where it’s harder to predict usage. You might not want to pay up-front for an entire server, or a year’s worth of use in advance, just to run a job that you know you might not be running 24 hours a day, 365 days a year.

In this scenario you might prefer to run virtual machines with a provider instead. On a per-unit-of-compute basis, you might pay a little more, but that is offset by the ability to switch it on or off more easily.

You likely still are responsible for some setup and configuration, but you might not need to lose sleep about being the poor soul who has to replace physical hardware if something breaks.

And if we compare to our earlier energy post, we might think of the properties here as something a bit like a coal fired power station compared to a nuclear one; where a generator is paying more per unit of electricity, but is able to ramp production up and down with ease.

Just like coal, virtual machines are now ubiquitous. But even in 2021, if you’re looking for virtual machines where all the energy used is matched by power coming from fossil free sources, there are fewer providers of this kind of compute than you’d think who will sell it at a more granular resolution than paying on per-month basis.

Managed platforms: hosted kubernetes or platform-as-a-service offerings like Heroku
Going further along the abstraction scale, rather than paying for a virtual machine that you set up yourself, you might accept a set of defaults provided by a platform which abstracts the underlying compute away from you even further.

Again, you’ll likely pay more per unit of compute, but the added abstraction makes it easier to ramp up and down. Also, increasing amounts of busy work that you might have done before is abstracted away too. And there is a greater degree of tolerance built into the platform to the underlying hardware failing.

The poster child of this kind of platform is a service called Heroku, but increasingly, an open source project called Kubernetes has grown to prominence as a basis for building many other more abstracted services like this.

Scaling down as well as scaling up
There are definitely benefits here. If you’re running services that are mainly used during office hours, then it’s likely they’re being used 40 hours out of 168 possible hours each week. This means that rather than paying for the three quarters of the time when they’re barely being used, you have the option of not running these machines.

Also, the same features that make a system more tolerant of hardware failure can make it handle scaling downwards as well as scaling upwards. It gives you a degree of flexibility that you didn’t have before, about when and how you run a workload.

This means it’s increasingly common to pay for this on an hourly basis and for it to be realistic to adjust the amount of computing capacity you’re paying for on an hourly basis when you’re working at this level of abstraction.

You don’t burn cycles needlessly; you don’t burn cash needlessly; and hopefully, it results in burning less fossil fuels too.

However, because these are often built on top of virtual machines themselves, you end up with even fewer providers to choose from, especially if you care about running them with cleaner energy. In many cases, if you want this level of abstraction, you might need to compromise on who the underlying providers are, or be explicit about choosing a greener provider in the supply chain.

Functions as a service or serverless: even higher unit cost, even more abstract, scaling built in
Finally, if we’re prepared to abstract away the underlying devices even further, you can end up with a model where you basically pay per invocation of a given workload, rather than paying to reserve a bunch of machine time in which you can run that workload.

If you run a website, instead of paying to run a webserver continuously, and hope that it can handle enough of the incoming from users, you might pay on a per request basis, and rely on the platform to take care of ramping up resources, and then ramping them down again.

It may be the case that what compute you end up consuming is even more expensive on a per-use basis than all the other options we’ve covered. But a bit like the gas turbines from our earlier energy example that we can start and stop easily, if you’re only ever paying for what you use, you can end up with very low total costs just because your total usage is so low. You are no longer paying to leave idling computing processes running, waiting for input between requests on a website.

If you only have a single, minute-long compute job to run at midnight, you probably don’t want to keep that machine running the other 23 hours and 59 minutes every day if you can help it.

Visualising a demand curve for compute

This is how I visualise the different ways you might run a single computing job these days: I picture a demand curve just like how we might picture one in the energy sector, and the myriad ways to generate electricity.

I hope it shows that some ideas from other domains can help us think differently about digital infrastructure. Because once you start thinking of it like an essential utility our society is built on, you can bring to bear all kinds of other ideas.

What this has to do with a greener digital sector
Getting back to the subject of compute powered by green energy, if we know we are able to think independently about the kind of computing job we want to run, and where or how we want to run it, it allows for re-thinking what a datacentre should look like, or where we want to build reliability into any system we use.

An interesting example

One interesting example here would be the German company, Examesh. They build datacentres into the base of wind turbines, to take advantage of power that would otherwise be wasted, or to use the energy jargon, curtailed.

We say energy is curtailed when it can’t be fed into the grid, and put to productive use. This often happens because there is no way to move the power to where a workload needs to be run, due to a local oversupply of power on the grid causing ‘congestion’ on the power lines.

But if you can move the workload to where the power is instead, you can change the economics, and the clean, variable power that would be wasted can be put to productive use instead. Better yet, if we’re able to make use of existing physical infrastructure like this, we can avoid building another purpose-built datacentre building just to service those workloads.

This is a different model to the more common approach of building massive datacentres which stand apart from the rest of the built environment. And it’s partly made possible by us thinking about compute workloads as things we can run on different kinds of fungible sources of computing power–and realising that by matching them to different sources, we can optimise for cost, for carbon emissions, speed, and so on.

This isn’t the only idea from the power sector that can help us think differently about digital infrastructure, but I do think it is a useful one. We will explore some more of these ideas in future posts.

Originally published in the The Green Web Foundation

Meet Our Steering Committee: NTT DATA

Green Software Foundation — Wed, 02 Feb 2022 23:18:23 +0000

NTT DATA on green software and the challenge of striking a balance between introducing technologies that enrich society and reducing CO2 emissions and the environmental impact.

NTT DATA—a part of NTT Group— is a trusted global innovator of IT and business services headquartered in Tokyo. We help clients transform through consulting, industry solutions, business process services, IT modernization, and managed services. NTT DATA enables clients as well as society to move confidently into the digital future. We are committed to our client's long-term success and combine global reach with local client attention to serve them with more than 130,000 employees in over 50 countries.

What does NTT DATA plan to achieve together with the Foundation, for your business and for the world?
As a part of the NTT group, NTT DATA aims to achieve zero environmental impact following the vision, "NTT Green Innovation toward 2040".

Since the importance of IT systems will continue to increase throughout society, we will take responsibility as an IT service provider and work to reduce carbon emissions in the development and operation of software and systems themselves. In system development, we will work to establish new emission standards that reflect our efforts to reduce carbon emissions and promote initiatives to change the awareness of system development within our own company and society as a whole. Through these activities, we will contribute to the primary goal of reducing CO2 emissions by 45% worldwide as an IT service provider and aim to realize a sustainable society where the environment and IT systems are in harmony.

How does NTT DATA see the future of green software?
While digitalization is accelerating and society is becoming more affluent, the scope and role of software in society will continue to expand. At the same time, carbon emissions from the development and operation of such software will continue to increase, leading to a less sustainable society. NTT DATA believes that it is essential to reduce emissions further and improve power usage efficiency by working on technological innovation in the software that determines the efficiency of the hardware, networks, and data centers. NTT DATA believes it is essential to promote a carbon-aware software development style that effectively incorporates these technologies. Through the GSF activities, NTT DATA aims to make software greener by developing technologies to properly understand and reduce carbon emissions from software use, developing integrated development methodologies, and realizing them as practical tools.

What challenges can get in the way of progress of green software?
As represented by the ever-evolving AI technology, new technologies are being developed one after another in the IT field to enrich society and help solve social problems. On the other hand, these technologies' research, development, and diffusion may increase carbon emissions. The challenge for the software field will be to strike a balance between introducing these technologies that enrich society and reducing CO2 emissions and the environmental impact. We should solve these issues by discussing them with stakeholders, introducing new technologies, and building an eco-system that harmonizes the two.

Meet the other Steering Committee Members of Green Software Foundation:
Read about our Steering Committee members, what they say about working with the Foundation and their thoughts on the future of green software.

Accenture | GitHub | Globant | Microsoft | Thoughtworks

Originally published in the Green Software Foundation

Meet GSF Org Leads: Colleen Josephson of VMware

Green Software Foundation — Wed, 02 Feb 2022 23:08:05 +0000

Colleen, a Research Scientist, explains how the VMware team’s interest comes from a passion for doing what’s right, a willingness to tackle hard problems, and the belief that doing good for the planet aligns strongly with the best interests of their customers.

First tell us about VMware and what your organization expects to achieve from the association with GSF.
VMware is the essential software element that powers most mission critical business systems today. The software that runs the world runs on VMware compute, cloud, mobility, networking and security offerings. We believe that software has the power to unlock new opportunities for people and our planet. We deeply value sustainability and it is an important tenet of our ESG 2030 Agenda.

We joined the GSF because we seek to both inspire and implement the next generation of sustainable digital infrastructure. We want to be a part of creating tools and processes to facilitate greener, more sustainable software engineering.

Tell us about you and your team’s interests and VMware’s journey to GSF.
I myself am a research scientist (Distributed Edge, OCTO) who focuses on sustainable computing. But it's hard to talk about this work from an individual perspective because VMware’s efforts in green software really are a team effort. Our journey to the GSF wouldn’t have been possible without Zhelong Pan (Senior Architect, Sustainable Software Solutions), Andrew Woosnam (Software Engineer, Modern Applications Platform), Jeremy Wilken (Clarity Design System Manager, Design Platform), Nicola Peill-Moelter (Director of Sustainability Innovation) and Shalini Singh (Senior Manager, Sustainability Operations).

Some of us have ‘sustainability’ in our job titles, but for all of us, our interest comes from a passion for doing what is right, willingness to tackle hard problems, and a belief that doing good for the planet aligns strongly with doing good for our customers.

How do you view working with the GSF and in green software in general?

Sustainability and green software are big problems that no one company can solve, so we need this space where lots of contributors can get together and exchange ideas on really hard problems. I really value the community of engineers who care deeply about sustainability. It is an enormous opportunity to learn from people who have expertise in all sorts of areas, all of us working together towards a common goal.

What obstacles do you see to wider use of green software? How do you think we can overcome them?
I see the biggest obstacles as culture and tools. Only recently have we started talking about the energy aspects of software. Far more often we hear about other metrics like computational or space complexity. We teach those concepts in CS classes, we ask about them in job interviews. So naturally software engineers infer that these matter most.

We need to change what we’re talking about, have more conversations about things like energy and sustainability related metrics. I think the cultural challenge plays a big part in the second obstacle: lack of tools.

Right now, we have a lot of tools to help us analyze memory usage, code duplication, concurrency, and so on. We have very few to help us analyze energy consumption and carbon intensity. Without tools to help them, even the most enthusiastic software engineer will have trouble actually implementing greener software at scale.

What other matters would you like to share with our readers?
Your readers may want to check out these links to VMware’s relevant work in the sustainability space:

The Path Toward Sustainable 6G

Pathways to Sustainable Clouds

Accelerating Decarbonization of the World’s Digital Infrastructure

This article is licenced under Creative Commons (CC BY 4.0)

Want to know more? Sign up here.

MEET GSF ORG LEADS: TIMO MULLER OF LEADERS FOR CLIMATE ACTION

Green Software Foundation — Wed, 02 Feb 2022 22:54:09 +0000

In this second interview in the GSF Org Leads series, Timo Muller of Leaders for Climate Action (LFCA) talks to us about how joining hands with Green Software Foundation helps spread the message of green software more effectively.

Tell us about LFCA and what your organization expects to achieve from the association with the Green Software Foundation.
Leaders for Climate Action is a community of entrepreneurs and their companies uniting to drive impactful climate action in their organisations and beyond. Today we are more than 1500 companies and are operating on a global scale. Our members commit to ambitious climate goals and we provide them with everything they need to achieve these goals.

We look at climate action through a community lense. We don't provide individual consulting to companies. Instead, we split our community into sectors, provide these with concrete action items on our platform and help them through the implementation phase by building small mastermind groups with peers. Additionally we drive campaigns and speak up on a political level as representatives of a forward thinking industry.

We work with a lot of companies from the digital sector and are therefore very aware of the specific challenges that we are facing as an industry. The high demand and corresponding growth of digital services is causing an ever-increasing footprint of our economy.

By cooperating with the Green Software Foundation we hope to accelerate the transition towards a digital economy that runs entirely on renewable energy. As a player between the private and public sector, we can help to increase adaption of new technology in the field and are already working on a few interesting initiatives in this space. While we see the challenges, we also see our industry as a potential frontrunner and enabler of the much needed broad transition of our economies.

Tell us about your career path, interest in green software and your journey to GSF.
I studied computer science and have been building digital products ever since I learned how to use a computer. I dropped out of university to found my first startup. After living through all the ups and downs of the startup world, we sold our company to a competitor. I have worked there for 2 years as their Head of Product.

During those years I started realising that we are facing an environmental crisis that is far too big to ignore. I began to question my personal lifestyle choices, stopped eating meat, reduced flying and started investing in renewable energies. This journey finally led to me and my former co-founder Philippe founding LFCA, which was an idea of one of our previous investors.

I am committed to dedicating my professional career and skills to fight the climate crisis. I love coding and using tech to solve challenges, because it enables individuals and small teams to achieve change on an incredible scale. We need to leverage the creativity of the global tech community for the challenge ahead. But in order for us to do this, we need standards and open APIs that bridge the software to the energy layer. The GSF can play a decisive role in achieving this.

I met Asim (Chair of the Green Software Foundation) through the CAT (Climate Action Tech) community quite a while ago when LFCA was still in its infancy. We have been in contact ever since and joined the GSF as a founding member.

What do you, as an individual, expect to achieve by working with the GSF and in green software in general?
As an engineer, I expect to test, and if time permits, also contribute to the products that the GSF is working on. Building truly carbon aware software is exciting and I am looking forward to being an early adopter of new tech in this field. Additionally, I expect that we, as an organization, can help distribute new standards in the tech community and convince companies to adopt these.

What obstacles do you see to popularising green software? How do you think we can overcome them?
The reality for CTOs of big corporations is that green software is not their first priority. Stability, security and speed of development will come first in most cases, and we should not expect this to change. Instead, we should make sure that the solutions we provide are reliable, well tested and don't need much management overhead. The fact that many tech giants are on the forefront of this development makes me optimistic that this will be the case in the near future. It's likely that the cloud providers will provide some abstraction layer at some point. Organisations like the GSF and LFCA are essential in increasing the pressure for this to happen sooner rather than later.

When it comes to adoption in the tech and open source community, I do believe that it would be beneficial to get some of the thought leaders in this space on board. The Open Source community is relatively small, with a reasonable amount of "influencers" that are often active on Twitter. Communities are a social phenomenon and often enough things become "cool" just because some people started adopting a certain kind of technology. It might make sense to invest time in getting some of these thought leaders on board and ask them to start talking about green software. Of course, this first requires an initial set of tooling for people to start experimenting with.

Any other matters you would like to share with us?
We are launching our next "Time for Climate Action" campaign around Earth Day 2022. If you want your company to participate, please reach out.

This article is licenced under Creative Commons (CC BY 4.0)

Want to know more? Sign up here.

MEET A STEERING COMMITTEE MEMBER: DAN LEWIS-TOAKLEY OF THOUGHTWORKS

Green Software Foundation — Fri, 28 Jan 2022 17:41:55 +0000

Tell us about your career path, interest in green software and your journey to GSF
As a technologist, I’m passionate about green software because I want us together to solve the climate crisis. As the demand for digital products and services grows—especially during the era of COVID-19—we have an increasing responsibility to develop software with limited or no reliance on fossil fuels.

As the Green Cloud Lead for Thoughtworks North America, I advise clients on strategies to help reduce the environmental footprint of their technology. I also work with climate justice movements to deliver software solutions for increased growth and effectiveness. Prior to Thoughtworks, I spent almost a decade working with nonprofits to build technology solutions to scale their impact on issues of climate, social and economic justice.

What do you, as an individual, expect to achieve by working with the GSF and in green software?
Ultimately, I want to see us build a large and diverse movement of green software practitioners all around the world. More specifically, I’m hoping that we can innovate with new open source and open data projects to build a trusted and shared ecosystem of tools and techniques to build greener software. In particular I’m passionate about pushing the limits of what’s possible when it comes to architectural approaches to greener cloud computing, video/audio streaming and edge computing.

What obstacles do you see to the cause of green software? How can we overcome them?
I think one of the biggest challenges we face is time. If we want a critical mass of green software adoption over the coming years, we are going to need policy makers to set requirements and standards for doing so. This is a huge task that will require stakeholders across all industries and sectors to come together fast enough, while the impending climate crisis only grows in size and reach. I’m hopeful that we can overcome this challenge—but we need to move quickly.

Any other matters you would like to share with us in the cause of green software?
Related to green software (and a shameless plug)—our team at Thoughtworks has been building an open source tool to help measure and reduce the carbon emissions from cloud services and usage. It’s called Cloud Carbon Footprint - you should check it out and give it a try!

Check out what our other Steering Committee members say about green software
Jeff Sandquist of Microsoft

Elise Zelechowski of Thoughtworks

Erica Brescia of GitHub

Santiago Fontanarrosa of Globant

Toru Shimogaki of NTT DATA

This article is licenced under Creative Commons (CC BY 4.0)

MEET A STEERING COMMITTEE MEMBER: TORU SHIMOGAKI OF NTT DATA

Green Software Foundation — Fri, 28 Jan 2022 17:29:43 +0000

Meet Toru Shimogaki, the Head of Green Innovation Office and Head of Advanced Computing Technology Center at NTT DATA and learn what he thinks about green software.

Tell us about your career path, interest in green software and your journey to the Green Software Foundation (GSF)
As a software engineer, I initially worked on performance improvements and feature additions for PostgreSQL, an open-source software - RDBMS. I am the original developer of a data loading utility called pg_bulkload. After that, I supported the introduction of Apache Hadoop, Apache Spark, and Apache Kafka, which are big data platforms.

I also wrote four books with my colleague on the subject and worked on dissemination and deployment.

*What do you, as an individual, expect to achieve by working with the GSF and in green software? *
When it comes to CO2 reduction in IT, it seems that the focus is on hardware and data centers. However, as a member of the software industry, I realized through this project that software used by hardware and data centers has a great power to contribute to climate change issues. The issue of climate change is now a topic that everyone in the world should be aware of. I want to convey the message to people involved in software to be aware that they are not an exception.

What obstacles do you see to the cause of green software? How can we overcome them?
The first obstacle is that it is not intuitive and difficult to understand how software can contribute to CO2 emission reduction. One approach would be to develop emission standards for software, clarify how to use those standards, and make them easy to use.

In addition to improving the software itself, I think it is also necessary to develop easy-to-understand standards and methodologies for using the software for Carbon Aware.

Any other matters you would like to share with us in the cause of green software?
As someone who has dealt with big data infrastructure, I have been using computational resources without awareness. However, we should also look for ways to do the same processing with less environmental impact. Although promoting research needs many computational resources, we need to make efforts and be aware of both the social significance and the necessary research results.

Check out what our other Steering Committee members say about green software
Jeff Sandquist of Microsoft

Elise Zelechowski of Thoughtworks

Erica Brescia of GitHub

Santiago Fontanarrosa of Globant

Dan Lewis-Toakley of Thoughtworks

This article is licenced under Creative Commons (CC BY 4.0)

SUSTAINABILITY SUPERHEROES – CIOS ARE PULLING ON THEIR GREEN CAPES

Green Software Foundation — Fri, 28 Jan 2022 17:29:34 +0000

CIOs are leading the way in rethinking sustainability strategy as a majority of companies agree that sustainability is key to future performance and investments in digital technologies are on the rise to help companies achieve their sustainability objectives.

We know that sustainability is rapidly rising up the agenda on both a personal, and a professional level. Whether it’s home renewables and efficiencies, electric vehicles, growing consumer demand for more sustainable products, or corporate sustainability initiatives, the human interest is driving organizations to rethink their approach to sustainability. A recent blog discussing the journey beyond net zero highlighted Accenture research findings that 99% of CEOs from billion-plus dollar companies agree sustainability will be a key to future performance.

Accenture research also tells us that 75% of companies believe their investments in digital technologies will help them achieve their sustainability objectives.)

The role of the CIO must not be underestimated. In fact, IT departments can again be the changemakers, superheroes of the current green revolution, as they lead from the front to enable organizational sustainability. And there’s an obvious place to start.

1. Cloud migrations cut carbon
It’s pretty obvious the data center is carbon intensive with all its power requirements, and it’s clearly something of an anomaly in an environmentally conscious world. When you go to bed you don’t leave the lights on all night, you simply switch a light on when you need it. In commercial buildings it’s common for lights to be on sensors – they’re only on if people are moving. And it’s the same when you think about the datacenter – if a resource is not in use there’s no need for it to run. Why should it be any different for servers?

We know that most organizations have made some moves to the cloud, but did you know that moving to the Microsoft cloud can lower your carbon footprint by up to 98%? And, with Microsoft’s commitment to being 100% renewable powered by 2025, and carbon negative by 2030, the sustainability benefits speak for themselves. And of course, the more you use Azure,the greater the potential for reducing your carbon footprint.

But let’s not pretend that simply moving to the cloud is enough, that’s just the start. 

2. The power of the cloud enables greater change
The cloud enables an organization to measure, monitor and act, to become future ready by enabling the business to take action through support, tracking and empowerment. The first place to look is cost and consumption. While Azure will be powered by 100% renewable energy by 2025, it’s best practice to optimize the cloud for cost, which is tightly coupled to energy usage. If you’re mismanaging costs, you’re likely mismanaging carbon. And consider another core part of sustainability, recycling; if resources are running but unused, they should be reallocated – reducing wastage is critical to cloud optimization, and sustainability.

Beyond cloud migration, consumption and efficiency, the incredible power of Azure provides organizations with ways to not only collect and store carbon related data, but more importantly, the ability to uncover insights, ideas and recommendations. And, the power to react quickly and at scale, through automation. Imagine for example, the ways in which you could analyze data to optimize your logistics and bring carbon efficiencies to your fleet, the dashboards you could create for facilities teams to track and improve energy use across your buildings. Technology can be the starting point for an ecosystem of sustainability opportunities. 

3. You need a rounded sustainability strategy
You’ll already have dashboards within your business, but do you have one for sustainability? Many of those dashboards can be viewed through a sustainability lens, but without a dedicated, focused view, any strategies you do have will be lacking. If you’re not accurately measuring, you’re not going to drive significant change. But a note of caution is due here – there’s nothing more frustrating than seeing an issue and being unable to act. It’s key to empower for accountability, but critically, organizations must empower their people to act.

Some organizations will have a Chief Sustainability Officer, some will have sustainability as an additional executive responsibility, while others will not have determined who is responsible. Whatever the scenario in a business, the CIO will have a fundamental role in developing, enabling and driving the strategy. Whoever owns the strategy, it must cover Scope 1-3 emissions (emissions from owned sources, emissions from purchased energy and all other emissions), and progress must be carefully tracked.

For long term success and to drive real change, sustainability needs to be embedded in the very fabric of the organization. A mantra of sustainable by design must be embraced, where anything new is designed to be sustainable, and every employee is a sustainability champion, questioning wastage and asking why dev machines or test instances are not automatically shut down when not in use. Operations must by default become sustainable, with decisions viewed through a new lens, this is not a one and done scenario. And you need KPIs, so start tracking your data points now. When it comes to measuring progress, you’ll be thankful you created a starting point early.

The time to act is now
With the recent announcement of the Microsoft cloud for sustainability, organizations will be able to discover and connect to real-time data sources to uncover actionable insights. By aggregating sustainability and enabling organizations to simply record, report, reduce and replace their emissions, now is the time to elevate the sustainability agenda.

Originally published in the Avanade Insights Blog

Want to know more? Sign up to our newsletter here.

MEET GSF ORG LEADS: ZILIANG ZONG OF TEXAS STATE UNIVERSITY

Green Software Foundation — Thu, 27 Jan 2022 10:03:15 +0000

Tell us about your work at the Energy Efficient Computing and Systems Laboratory at Texas State University and your work on green software education.
Texas State University (TXST) is fully committed to and has a long history to promote the research and education in sustainability in general and sustainable computing in particular. As the director of the Energy Efficient Computing and Systems Laboratory at TXST, I have been fortunate to work with a great team of colleagues from both industry and academia, and many talented students on addressing challenging problems in green software engineering, green AI, green data center, and green cloud with nearly 70 publications in these fields.

We developed the Marcher system, which has successfully supported a group of researchers to publish over 20 papers.

As for education, I noticed many years ago that very few universities introduce "green computing" topics in the core CS curriculum. This worries me because hundreds of thousands of CS students are graduated every year worldwide without the "green thinking" mindset and proper training. They will be the next generation of software engineers who write millions of lines of code throughout their careers.

To address this problem, I created and taught three green computing courses (at the undergraduate, master and PhD level respectively) at the CS Department of TXST. Approximately 200 students have learned this topic from the courses I taught in the past 8 years.[Check out this 2021 paper on Bringing Green Software to Computer Science Curriculum: Perspectives from Researchers and Educators for illumination on how green topics are covered (and not) in CS curricula.]

In addition, we developed the GreenCode platform to support other educators who want to teach green software design, which has been used by professors from the U.S., Asia, and Europe in various computer science courses and received over 31,000 code submissions to date.

Tell us about your career path, your interest in green software and your journey to GSF.
I received my PhD degree from Auburn University, where I studied energy efficient resource management for high performance computing systems. About 15 to 20 years ago, energy efficiency was primarily discussed in the community of national labs and large data centers. Since 2007, we have witnessed the booming of mobile computing led by the revolutionary innovation of the iPhone. Due to limited battery life, energy efficiency gradually becomes a factor that software developers need to consider.

I started to expand my research on improving the energy efficiency of general software after I joined Texas State University in 2011. Recently, with the large deployment of AI, Blockchain, VR/AR, and Cloud applications, environmental impact has grown to be a first-class concern that requires the full attention and close collaboration of all stakeholders in the software community. To share our previous research/education experiences with a broader audience and contribute more to the community, I joined the Green Software Foundation (GSF) and the work I have done with GSF colleagues in the past months has been exciting.

What do you, as an individual, expect to achieve by working with the GSF and in green software in general?
I want to work with the GSF community and contribute on developing power measurement tools that support green software design and education, making metrics for green software evaluation, finding solutions for challenging research problems in the field, and forming a software ecosystem that cares about greenhouse emissions and encourages creative thinking.

What obstacles do you see to popularising green software? How do you think we can overcome them?
The development of green software will encounter different obstacles at different stages. At the current early stage, the major obstacles are the lack of metrics, tooling, and awareness. To overcome this, we need to have a widely acceptable metric in place to evaluate the carbon impact of different software and develop urgently needed measurement and monitoring tools for software developers to practice green software design. We must also increase the awareness of green software through higher education, enterprise training, and social outreach.

Any other thoughts on green software and sustainability?
Software is eating the world and gets bloated faster. As hardware improvement quickly slows down in the post Moore era, it becomes the responsibility and inevitable mission of the software community to take the lead in advancing the frontier of green computing. Time is running out and bold actions must be taken now. I believe we will have a promising future if all stakeholders from industry and academia can work closely and make a long-term effort together.

Want to know more? Sign up to our newsletter here.

This article is licenced under Creative Commons (CC BY 4.0)

Sustainable Tech Choices for Cloud

Green Software Foundation — Wed, 26 Jan 2022 23:06:07 +0000

Running highly available products and services on top of variably available power calls for flexibility and targeted efficiency from cloud providers and users.

Introduction
This is the first of a five-part series covering Sustainable Tech Choices for Cloud.

At the end of 2021, AWS added sustainability to the well-architected pillars that define their “key concepts, design principles, and architectural best practices for designing and running workloads in the cloud.” A year earlier, Azure had posted their sustainable architecture guide, written by our own Asim Hussain. The good news is, the stories are similar—on climate-friendly software, we all know what to do.

So, why have both of the biggest public cloud providers now come out with guides to writing green software? In effect, they’re saying that sustainability is a fundamental part of how we need to think about cloud engineering from now on. They’re right, and Amazon, Microsoft, and (eventually, we predict) Google saying it was inevitable. As soon as they made their 2020 promises to be carbon zero in their own data centers by 2030, a change to the way the rest of us architect for the cloud was locked in.

Cloud providers cannot achieve CZero on their own
The first hint that the big three hypercloud providers couldn’t achieve CZero without change from their users came in June 2021 in a paper about achieving carbon zero on the Google Cloud. To my reading, Google implied they weren’t currently attempting to achieve their goals in their public cloud because it was too tricky—customer workloads were too opaque to them.

However, whilst Google Cloud Platform (GCP) can clearly focus on their own private workloads first, AWS and Azure don’t have that luxury. Customer workloads are their business, particularly for AWS. They need to find a way to steer their users towards the green services and practices that will enable Amazon and Microsoft to run their data centers on, primarily, renewable power.

However, running highly available products and services on top of variably available (i.e. wind and solar) power requires new, clever flexibility and targeted efficiency. Not just from them. From their users too.

The approaches we believe they are (or will be) promoting are covered in a new whitepaper on sustainable architecture for the cloud from the Coed:Ethics community in London, which the GSF is closely connected to. The whitepaper came out in January 2022.

Discussion of these new approaches is desperately needed because enterprises must update their systems to handle a world of variably available power. It won’t be easy, but it will help save the planet, as well as reducing their hosting costs and safeguarding them from a future in which burning fossil fuels to power data centers is no longer an option.

In the next articles in this series, we will discuss the four areas covered by the paper:

Operational efficiency (i.e. improving how software is run).
Architecting for minimal carbon (improving how software is designed).
Hardware efficiency (in particular, improving how end user devices are managed).
Energy efficiency (minimization of CPU/GPU and network use).

Steps for enterprises that operate in the cloud
In summary, the paper recommends the following steps for enterprises that operate in the cloud:

Select or transition to hosting providers with a solid commitment to sustainable hosting. This means those that have a committed date for 100% sustainable power and are making clear progress towards it. And push them to move faster. Often the easiest way to do this is to state a strong interest in sustainability to whichever host you choose.
Host in regions that provide excellent options for sustainable electricity provision. Cloud providers should indicate which are their preferred regions for sustainable expansion.
Assess the hosting efficiency of cloud providers. You can do so by asking whether the current strategy for machine utilisation is good enough. Use high density managed services where possible.
Turn off unnecessary, unused workloads, introduce autoscaling, and rightsize machines.
Review architecture for expensive tasks that are on demand or always on but do not need to be. These are tasks that are expensive in terms of processing, but this often corresponds to $$$. Separate them out and use spot (AWS/Azure) or preemptible (GCP) instances to run them or flexible (e.g. burstable) instance types wherever possible.
Reduce the carbon impact of costly tasks that must be executed on demand. For example this can be achieved by using more lightweight languages, frameworks or services or by moving more work to the client or edge.
Have a ten year policy on user device longevity. And don’t let apps that the enterprise has created be a nail in the coffin of working hardware for end users.

Most importantly, enterprises need to measure carbon emissions—or their cost proxy where appropriate—and pick their battles. There is considerable work to do and it’s vital not to waste precious time and resources optimizing software that has little impact.

The Green Software Foundation’s Software Carbon Intensity specification is there to advocate for this. But tools already exist, for example, Google’s Carbon Footprint, Microsoft Emissions Impact Dashboard or the open source Cloud Carbon Footprint tool. An AWS one is coming in 2022. Cost can also be a rough guide to emissions, and AWS Cost Explorer or Azure’s cost analysis can help with this.

What Next?
The next article of this series will focus on operational efficiency in more detail - what it means, and what services exist to help improve it.

Want to know more? Sign up to our newsletter here.

This article is licenced under Creative Commons (CC BY 4.0)

MARCHER: A FINE GRAINED SOFTWARE ENERGY EFFICIENCY MEASURING PLATFORM

Green Software Foundation — Tue, 18 Jan 2022 14:56:58 +0000

Discover an easy to use set of tools for fine grain software energy profiling for developers, students and researchers.

Energy efficient computing is big business. With data centers now at 2% of all US energy consumption in the US, it is fair to say that cutting energy bills is a top priority for all major tech companies. Since 2011 we have seen admirable improvements in data center efficiency with consumption increases dropping from 25% from 2005-2010 to 4% from 2011-2014 currently. (Note that current trends depend on current throughput and consumption trends to stay consistent).

This trend has been driven largely by super-efficient, hyper-scaling centers from the likes of Google, Amazon, and Facebook, built from the ground up to be as thermally efficient as possible; as well as widespread industry adoption of server virtualization technology and better job schedulers. The low hanging fruit, however, will soon all be taken.

Server virtualization is here, and it is hard to imagine another magic bullet solution like it to come along soon, as latency and other demands place upper bounds on just how much utilization a server can reach for web applications. The number of companies that are able to migrate to platform as a service data centers are dwindling. And hardware manufacturers continue to struggle in their downhill fight against Moore’s law. In view of all this, we will have to look to new avenues to keep up with consumer demand.

Big data becomes huge data...
This period of 1-3% growth in yearly data center energy consumption could also come crashing to a halt as what we call big data becomes huge data. By 2025 we will be creating more data in one year than in all of previous recorded history. Quickly, 4k and 8k streaming may become the norm. And 5G networks will allow mobile users to consume more data than ever before. Power hungry deep neural networks will be rolled out to an increasingly wider range of applications.

If you are committed to energy efficient computing for the future and sustainable growth of infrastructure and capacity, it becomes clear that hardware and data center design alone will no longer be a long-term solution. Software efficiency must be a primary pillar in achieving these goals.

Software energy efficiency calls for measuring tools
To improve software energy efficiency, we would have to measure it accurately first. However, measuring software energy consumption is not an easy task. To understand energy characteristics, you need to know not only how much energy a program consumed, and how long it ran, but how each component was being used. These measurements should also be made as easy as possible and support a wide variety of tools and languages.

Enter The Marcher
All of these factors were the motivation for creating the Marcher system. We wanted to create a platform to help developers, researchers and students study the energy efficiency of their code. And we wanted the devices to be equipped to measure the kind of workloads seen in industry—such as for deep learning and big data processing—as well as in typical high performance computing systems like in the clustered parallel processing of batch jobs.

With funding from the National Science Foundation (NSF), we have built our custom systems with two Intel Xeon processors, a Nvidia GPU compatible with CUDA, and an Intel Xeon Phi coprocessor, 32 GB of DDR3 DRAM, one 160 GB SSD and one 1 TB HDD.

The system supports 20+ popular programming languages including C/C++, Python, Java, CUDA and C#.

To facilitate ease of access and use we have provided what we call Power Profiling as a Service (PPaaS) allowing users to submit and run code directly in-browser through GreenCode.

In addition to this cloud service, for processes which require multiple nodes or longer capture times, we have also built robust command line tools users can access over ssh after applying for credentials.

The power profiling of the CPUS, GPUs, and Xeon Phis are handled by the Intel RAPL interface, the NVIDIA NVML interface and the Intel MICAccessAPI respectively. For all other components that do not have built-in sensors—such as Solid State and Hard Drives—we developed external power sensors called Power Data Acquisition Cards (PODAC). This collection of profiling tools allow us to find in a fine granularity how much power a program is consuming from each component at any given time.

And here's the Marcher System.

Between the web and command line interfaces sits our custom power profiling and job scheduling layer which dispatches jobs, as well as collects, aggregates, performs calculations and saves power data.

Examples of measurements with the Marcher System
As mentioned earlier, we wanted these systems to be able to measure power for real-world use cases. Here are several examples of measurements taken with the Marcher system.

Example 1
Here we see power data obtained from a Xeon Phi coprocessor running the Single Source Shortest Paths (a.ka. Dijkstra’s algorithm) OpenMP program on a single Marcher node.

Example 2
The following graph shows the power data collected on two marcher nodes running an MPI-2007 benchmark application.

Example 3
This graph plots the real-time power consumption data of CPU, DRAM, and GPU when training the AlexNet with Caffe on a Marcher server.

Marcher provides easy to use tools for fine grain software energy profiling, which empowers developers, students, and researchers to learn more about the energy consumption of their software. It provides Power Profiling as a Service through both web-based and command line interfaces. If you would like to read more details about the development of the Marcher system, read our paper here.

Originally published in the GreenSoft, Texas State University

Want to know more? Sign up here.

MEET OUR STEERING COMMITTEE: GITHUB

Green Software Foundation — Tue, 18 Jan 2022 11:06:47 +0000

GitHub shares thoughts on green software and how developers have to make sustainability-conscious decisions throughout the process of building and shipping software.

GitHub is the home for all developers. We make it easier for developers to be developers: to work together, to solve challenging problems, and to create the world’s most important technologies. We foster a collaborative community that can come together—as individuals and in teams—to create the future of software and make a difference in the world.

GitHub is much more than code. It is the home of open source collaboration. It is where new developers get started and where experienced developers expand their knowledge. It is a community where developers come together to solve the unsolvable and test the limits of what software can do. Our community is made up of students, hobbyists, consultants, enterprise professionals, partners and executives, building software in the way that works best for them.

What does GitHub plan to achieve together with the Foundation, for your business and for the world?
Climate change and the depletion of our natural resources are such large and vast problems that no one person can solve them. It requires a collective effort. The developer community on GitHub is diverse and collaborative, and together, they have the power to create solutions that accelerate human progress and positively impact our planet. We’re committed to building an environmentally sustainable home for all developers so they can build software that improves the health of our planet.

We’re also working to make carbon-free software a standard across industries. GitHub has been carbon neutral since 2019 and all development on and use of GitHub.com since then has been carbon neutral. To create greener software, developers have to make sustainability-conscious decisions throughout the process of building and shipping software. Creating standards in partnership with the Green Software Foundation will make it possible to increase the adoption of green practices in software development.

And finally, we’ve created a program, GitHub Sponsors, that allows others to sponsor contributors and maintainers working on projects that they’re passionate about. There are many green-software-focused projects built on GitHub and we’re excited to explore how we can deliver funding to developers creating and advancing green software initiatives through the GitHub Sponsors program.

How does GitHub see the future of green software?
According to NASA, Earth's temperature has already risen by over 1 degree centigrade since the 19th century, and carbon dioxide levels in the atmosphere are the highest they have been in a millennium. We must take urgent action to bring down emissions and mitigate the future impacts of climate change.

Every line of code has an environmental impact, with the resulting software consuming earth’s resources. Burgeoning technologies like video streaming, cloud storage, cryptocurrency mining, and artificial intelligence require increased computing resources, and therefore, generate increased carbon emissions. The cycle of new technology isn’t going to slow any time soon, so it’s imperative that we change how technology is built.

We see a future where every software developer can incorporate sustainable-forward thinking into their workflows, even if they aren’t familiar with carbon-aware software development. By making code and platform architecture more efficient, developers have the power to reduce the carbon footprint of the machines running it.

What challenges can get in the way of progress?
The concept of sustainability means something different to different people, and that presents a big challenge. How do we standardize implementing sustainable-forward thinking into the software development lifecycle? How do we get every company across every industry to engage in creating green software at a deep enough level to make an impact? Many of these challenges can be solved through education and the work of passionate developers in the community practicing this in the companies they work for.

Accenture | Globant | Microsoft | NTT Data | Thoughtworks

Originally published in the Green Software Foundation blog

Want to know more? Sign up here.