I started working with technology as an intern in India for a small startup which was building software for Financial Services companies. The network architecture that we were building then was client server, for those who are wondering what the heck is that click on the link. The servers were just another personal computer that was running the database application and the client was a proprietary software that was build onto of the database vendors technologies. The popular ones then were dBase III, FoxPro, Visual FoxPro, Oracle+Pro C, Lotus Notes etc I learnt a lot about how this technology worked and how this connected architecture delivered value to our clients. It was quirky, but did the job most of the time and we were a startup so we jumped off the cliff and built a parachute on our way down… for the most past it was a lot of fun and a lot of hard work. Fast forward to now, the technology landscape has dramatically changed. Internet and Mobile access has fundamentally changed everything about technology application in business.
While we embark on this new journey, I feel the technology community is not doing a good job of communicating the challenges of technology to the main stream. Lets start with the use of data centers, most of the worlds largest data centers get their electricity from non-renewable power i.e. Coal, Oil or Nuclear. That is a fact and there is nothing wrong with that but this has serious implications for new hyper connected, big data, high performance computing requirements of today’s consumers and the applications they use. There are billions of devices connected to the internet and all of them creating, storing, processing data for various things that puts CO2 into our atmosphere. One can debate until the cows go home whether CO2 in the atmosphere is a bad thing I am not that intelligent, it is accepted that CO2 in the atmosphere is a bad thing and it causes climate change. It is estimated that 2% of the CO2 in the atmosphere is due to data centers. This trend is serious because if one measured the CO2 due to data centers 10 years back it would have insignificant and McKinsey estimates that by 2020 4% of the CO2 will be due data centers. Again doubling in just a decade, why is this important or bad… it goes back to the notion of geometric progressions, we humans are very bad at calculating the consequences of exponential or Power Law based systems. We mess up with risk calculations in financial markets because we don’t understand the convexity of risk in high order moments, but that has only economic impact, painful as it may be we can cope with it and recover, but large complex systems like climate are too serious a threat to be left to Governments to bail us out.
What are the industry leaders like Google, Amazon, Apple, Microsoft etc doing about this? They are changing the debate from Sustainability to one of efficiency. They argue that by making their data centers more efficient they are reducing the amount of energy used by them. Efficiency is a good thing but data centers becoming efficient does not reduce the CO2 emissions it actually accelerates it because the cost of using data centers becoming really cheap so it gives incentives for more use rather than less. This is a troubling trend. I have been a vocal about my dis-belief in setting up data centers in Iceland but when I see the arguments and reports and lobby coming out of the large enterprises I am compelled to reject that notion and I want to jump in and change the debate.
I have accepted a new role of CEO of GreenQloud of Iceland, we are going to change the debate. We are vehemently going to argue that Efficiency != Sustainability, we believe data centers should be using only renewable power and should not be buying power from dirty sources. That is the only way to sustainability. GreenQloud is the Worlds First Truly Green(tm) Public Compute Cloud Service. I want to make a truly world class compute cloud that does not charge a green premium, is Amazon compatible and it provides the same or better performance as any other cloud computing platform in the world and is also sustainable. I believe that is the only way to disrupt the trend that we are on. We want to make the choice to choose sustainability easier, we don’t believe we should expect companies to sacrifice anything (read pay more!) to choose sustainability.
I will continue to blog in the Startup Iceland blog about Entrepreneurship, Startups and resources for entrepreneurs, now I will have more real world examples because GreenQloud is a startup, however I take the challenge of sustainability very seriously so I will be writing as well in the GreenQloud blog the emphasis there would be focused on what we are doing at GreenQloud. I will be recruiting other bloggers to write in the Startup Iceland blog because this is a platform that belongs to the Startup Community here in Iceland, I will continue to be a moderator and bar tender. Sustainability is a very powerful word, we will continue on that theme here in Startup Iceland. I will be sending out a small survey to all those who have shown interest and for those who participated in Startup Iceland 2012 to get feedback to improve the conference for next year.
- GreenQloud – A Startup Profile (startupiceland.com)
- Data Centers Power Consumption and CO2 Emissions (arnoudm.wordpress.com)
- Trends in Data Center Optimization (datacenterknowledge.com)
- HP: New data center technology goes easy on the grid (networkworld.com)
- That Boom You Hear Is the Cloud (sys-con.com)
- Is Cloud Computing a Green Giant? (sys-con.com)
Very good point regarding Efficiency vs. Sustainability – it is
counter-intuitive to so many people that making something more
efficient effectively makes it cheaper and cheaper things tend to get
used more, not the same or less. This is such a confusing concept to
so many people, that it would probably have been worth explaining it
in more detail. 🙂
There is another downside to efficiency: it is the enemy of
reliability, at least in the realm of computing and many other
“systems” as well. Tuning a system so it has minimal waste and maximum
output inevitably means it also has effectively zero spare capacity to
handle unexpected events – because maintaining spare capacity is
inherently wasteful and thus inefficient. But in a large enough
system, the unexpected happens *all the time*, so a perfectly
efficient system may in practice also be so unstable that it is
It is my (admittedly uneducated) intuition that this is actually one
of the main underlying causes of how unstable the global economy has
become. Things have simply become too efficient and there’s not enough
redundancy and spare capacity left in the system to accommodate the
inevitable “unexpected” events…
This relates to GreenQloud’s business, because like all SaaS/PaaS
providers (green or otherwise), your profit margins will be directly
linked to your efficiency. Striking a balance between efficiency and
providing reliable service is one of the main challenges your team
I beg to differ. Efficiency and sustainability are two sides of the same coin. Efficiency fosters sustainability and you can’t sustain with being efficient on your resources.
I am referring specifically to the data center industry and those running large data centers. The issue that I am arguing for is not that you should not be efficient but efficiency is NOT enough. Because, the rate of computing, network and storage resources being consumed by the world today is growing at an exponential rate, we have been very poor at understanding the consequences of geometrically progressive systems therefore when a system is growing at this rate we cannot rely just on efficiency, we need something better. We should fundamentally ask if we are doing the right thing. I believe every data center should be powered by sustainable renewable energy sources. The debate in this community is always to brush off the “green” folks… I am not a “green” fanatic but I think it is important enough that I want to commit my time and resources to fight this cause because we owe it to our children and our children’s children.
Let me respond to your last statement about Profit Margins… I don’t believe that. If Profit Margin is the only motive for me to do this I would have stayed in the Bank where our cost of capital was less than 1% and we charged our customers anywhere between 8% to 16% return on investment. I don’t want to be part of that system. I believe there is a better way, yes, Profit Margin is important but it is not the most important thing and you can build a company with other values that are far more important. We need to sustain the economic engine of a company so we have to have a margin but not at the cost of sustainability. We will build a world class Cloud based system and we will do this in a sustainable way. Leadership is the difference, I choose to lead by example and live by example. I believe we can return a terrific return on investment to our shareholders without throwing the baby with the bathwater. You ask how are we going to do that, well, actually the cost structure of operating a data center in Iceland is extremely cheap. The cost of renewable power is extremely cheap as well, we will innovate on the software layer where we bring a lot of efficiencies that will deliver better value to our customers and our shareholders. One of those points is a project that we are working on right now I cannot go into the details but lets just say that we are doing somethings with how we position our compute nodes, connect them with Infiniband and also use SSD for storage that delivers exceptional IO, I mean a factor improvement of 5 to 10 times that value we can translate to price that our customers are more than willing to pay. You ask how do we deliver better Profit Margin without focusing on Efficiency? well, we do it by focusing on Innovation… we do better with the limited resources. Is it hard? of course… will it fail? maybe… that is why i joined a startup and not apply for a job with Microsoft.
I believe we can get tremendous value out of software based innovation, we are lazy with software because it does things at a very fast pace but we don´t ask the question is it the most optimized way to code a solution. You are a hacker ask yourself that do you always design software that is super efficient? I never did and I have always been able to go back to a piece of code that I had written to improve on performance or maintainability or throughput usually by factors of 2 or more. I believe we can refactor code and make it do things a lot more efficiently than when we first write it. That is what we will do in GreenQloud. Why do I believe we can get better at this? because we are a startup and our survivability depends on it…
Pingback: Re-Reading List | Startup Iceland