This document discusses the philosophy and benefits of open source appropriate technology (OSAT) and how it relates to cloud computing. It notes that OSAT has its roots in the 1960s culture of freely sharing and collaborating on software. The open source model can drive sustainable development by enabling production and localization at low costs. Cloud computing provides infrastructure that levels barriers and allows access to information technology, improving standards of living. The future of cloud computing is seen as distributed and federated, relying on open source technologies like containers and identity federations.
3. What is OSAT
Appropriate Technology in the context of Open Source
A framework in which the benefits of the open source methodology are
applied to technology which is of social importance
4. What is OSAT
A way to look at do-it-yourself and self-sufficient technology. Freely
available and modifiable by anybody
5. What is Appropriate Technologies?
Any technology which has a positive environmental
impact or improves living standards
6. • Software developed during the 1960s and
1970s created in academic / corporate
laboratories by scientists and engineers
• ARPANET, built in 1969, linked hundreds of
universities, defense contractors and research
laboratories
• Enabled mass sharing and collaboration
among users
OSAT Philosophy Dates
Back to the 1960s
“Hacker culture” emerged from labs
“freely give and exchange software they had
written, to modify and build upon each other’s
software both individually and collaboratively,
and to freely give out their modifications in
turn”.
Source: http://evhippel.files.wordpress.com/2013/08/private-collective-model-os.pdf), 2004
7. Today open source is widely used among
major internet companies
Open Source defect rates are 50x to 150x lower
than proprietary software1
Proprietary = time extensive and expensive
Open Source = building on existing code -
quicker to market and cheaper
Open source benefits a fast moving and rapidly
growing industry!
Creating differentiation
1. Wired, 2004
WHY?
9. OSAT sharing ideas through cloud
Ideas and blueprints crafted
and collaborated in the
cloud
Shared in global online
communities
Everyone with access to the
internet gains access to
vital, life improving
technology
10. The UN have said that there is a direct link between
access to information technology and development1
1Annan, 2000
The importance of accessing
knowledge through IT
11. The Open Source model can act as a
driver of sustainable development
1.It enables production as well as consumption;
2.It enables localization for communities that do not have the
resources to tempt commercial developers to provide local
versions of their products;
3.It can be free as in "gratis" as well as free as in "libre", an
important consideration for developing communities.
12. What is key for the cloud enabled OSAT
development?
INTERNET ACCESS!
13. Cloud computing as a means to provide
much needed infrastructure
• Cloud among the most significant disruptive technologies over the next two decades
• Third-world cloud computing providers using cloud to enable IT services in countries that would have
traditionally lacked the resources for widespread deployment of IT services3
• Quick and affordable way to tap into IT
infrastructure1
• Level the playing field, as it breaks down
barriers to entry2
2,3Cloud computing and developing nations (Greengard, 2010); 4Cloud computing - The business perspective (Marston, et.al, 2010); 5UN Conference on Trade
and Development (UNCTAD), 2013
15. What is Cloud?
NIST:
Cloud computing is a model for enabling ubiquitous, convenient, on-demand
network access to a shared pool of configurable computing resources (e.g.,
networks, servers, storage, applications, and services) that can be rapidly
provisioned and released with minimal management effort or service provider
interaction. This cloud model is composed of five essential characteristics, three
service models, and four deployment models.
Source: http://csrc.nist.gov/publications/nistpubs/800-145/SP800-145.pdf
16. Cloud to date
Has developed to the present into a quite centralized
architecture
Very few dominant players
To-date dominated by economies of scale and current
virtualization technology
17. 4 Cloud Categories
Telcos
VPS
providers
AWS
Google
Microsoft
local global
Health care
Education
Governmen
t
Financial
Aviations
Automotive
horizontal
vertical
27. QStack
Best of breed open source project packaged together in an easily
installable and usable form
28. Deployment agnostic = it can scale
with the business
Private can become hybrid, hybrid can
become public
QStack = sustainability enabled by
default
Burstability can allow for environmental
sustainability
29. The backbone of open source is
collaboration and contribution
Strengthening the compatibility of
open source by testing it against
proprietary requirements.
Creditability
through contribution
46. “Everything CoreOS is building has
been inspired by the way Google runs
its data center infrastructure”
- Alex Polvi, CEO CoreOS
Source: http://www.datacenterknowledge.com/archives/2014/07/16/etcd-secret-sauce-googles-kubernetes-pivotals-cloud-foundry/
47. Anybody will be able to set up a
cloud
wherever on the planet
Hi my name is Tryggvi Larusson.
I’m going to talk about open source and cloud computing in a wider context, and a bit on the future of the cloud.
Just to give a little context, the I want to briefly explain what we do at GreenQloud.
GreenQloud is a company out of Iceland and we provide two main services.
The first one is a public cloud which is focused on sustainability and runs solely on renewable energy.
It is an IaaS cloud and offers the equivalents of AWS EC2 and S3.
The second one we call QStack which is a packaged version of our public cloud basically and is installable as a software stack for an on premise cloud, which can be private, public or hybrid.
I wan to start with by explaining the OSAT concept, which is short for Open Source Appropriate Technology. Appropriate technology itself is you could say a superset of the ideas behind the open source movement and is really about a way to look at technology from a self sufficient and sustainable perspective, and as you can see my talk precisely these points line very well with the open source movements objectives and end results.
So the OSAT movement is about looking at technology from a do-it-yourself perspective, to enable people to make use of technology through freedom of information and the ability to share and modify to improve peoples lives.
So OSAT is about looking at technology from a sustainability standpoint, which rhymes really well with the core fundamentals of the open source movement.
If we look back in history we see that the same fundamental ideas that developed technologies such as the internets precursor the original Arpanet and the original Unix operating system and related technologies that came from hacker culture. This culture had its influence from the academic environment where people were used to come up with ideas, share them and build upon the ideas of others freely.
It is no secret to anybody here that the quality of open source components is in many cases significantly higher than in closed source project, at least in very active open source communities and the cloud and internet market has of course depended upon open source software components to a large extent and also have been bit contributors of these projects.
If we look at the linux kernel its amazing to see how fare it has come in these past 20 years. The code has grown more than 50fold, which maybe does not come down to its quality but it shows the support and activity that the project has gotten. Nowadays Linux totally dominates the HPC market where it has over 80% marketshare in the biggest supercomputers in the world.
This is just to show that the internet and the cloud have become fundamental aspects of sharing ideas which is a core concept of the open source and appropriate technology movements.
And sharing ideas is of course a fundamental aspect in the development of societies.
So just to highlight then the Open Source movement can be a very important tool for sustainable development through a society both socially and economically.
Of course for this all to work with open source collaboration, you need Internet access which is still a concern in the developing world but we are getting there. Internet usage actually doubled between 2006 and 2011, from 18 to 35%, and that was as I said 3 years ago.
Linked to internet usage the cloud is opening up so many things for especially the developing world as it is
Now lets look at the cloud, the state of the market and the technology as it stands today.
When we talk about the cloud this definition from NIST, which is the US National Institute of Standards and Technology, which is most popular to use.
I don’t have to read through the whold definition, but the words that really stand out there are the shared pool and on-demand elements. From there you derive cost savings which are based on sharing resources and economies of scale, then the on-demand element can roughly be summed up in that the cloud provider provides API access to the resources so that they can be pretty much instantaneusly turned on and off.
When we look at the cloud market we can see that there is a real consolidation in the market. The cloud has really grown into very few central points, which is surprising to some extent.
One player is by far the biggest one on the market which is AWS, with about 40% of the marketshare of the IaaS market.
This development has been influenced by the current tecnology used in the cloud space, but like I will discuss later I think that is about to change with the radical shift towards containerization.
Largely KVM and Xen which are of course open source.
The cloud market has really just four qadrants where the global players are the horizontal ones and the local players cater to the local markets such as the telcos, but the interesting thing is that open source is everywhere in these markets, maybe to the least part still in the local markets.
The current big cloud players have massive datacenters and the vast majority of those run on non-renewable energy sources such as coal, gas and nuclear.
So these big datacenters are sucking vast amounts of energy, and the big cloud players like Amazon and Facebook have been criticized for using so much dirty energy and this is of course on a massive scale so significan amounts of carbon could be saved in the atmosphere if these datacenters would run on renewable energy.
I’ve drawn a picture of the geographical locations of the biggest player on the market, which you might guess is AWS.
You see that the locations are quite few so close to half of the cloud internet traffic is going to these few locations that are almost all concentrated in the western countries of the world.
There are a few open source components that are absolutely fundamental to the development of the cloud as we know it. We almost don’t need to mention the role of Linux in this
but the most fundamental pieces really are the open source hypervisors, xen and kvm.
These two have then spun off the next layer above which is the cloud computing stacks, so you are probably familiar with openstack and apache cloudstack, which propelled kind of the next generation of clouds which have been smaller local clouds and private clouds.
Then we have some fundamental components such as mysql and rabbitmq which are almost found in all cloud software stacks.
I want to talk a bit about how our product is based around open source components and how we utilize them to deliver value to our customers.
Pretty much every component of our QStack is open source, or is based on an open source framework.
On our font end web ui side we chose to develop it on a quite new framework called Meteor. This framework is built on node.js . We previusly used to develop our UI in a python framework called Django, but Meteor has enabled us to double or triple the productivity of development as it is very well suited for exactly what we need which is highly dynamic and interactive web user interfaces. So I would say that meteor is the next generation web frameworks, specifically geared towards interactive user interfaces, after the generation of MVC style frameworks like Ruby on Rails and Django.
We combine meteor with back and services like elasticsearch and logstash, so for instance in our infrastructure management we collect all log information into logstash, and collect usage infrormation into elasticsearch, which an the be used to query usage data and draw up the graphs that you see on the screen.
A cloud is really not a cloud without an API so that is what we provide in the form of AWS compatible EC2 and S3 interfaces.
So one thing we did is that we helped the CloudStack community build an AWS EC2 compatible interface to the compute cloud service. You can see a screenshot of that being used here and I am using with it a commanline tool from another open source project which is called Eucalyptus which implements the client side of the EC2 API.
As well on the object storage side we in our current beta version we have incorporated OpenStacks Swift storage platform as an integrated part of QStack. This is also compatible with AWS’s S3 object storage api and I show here another open source tool which is called s3cmd which can be used to make comman line operations on the service api endpoint.
Here is an additional list of the most important components in the QStack software product. This is actually just a short list as we have a lot of smaller components or frameworks that we use that aren’t listed here.
For instance we have here the hypervisor layer which it in our case by default KVM which is in turn part of the Linux kernel.
We also make quite extensive use of chef as a devops style tool to automate deployments and mange updates of the software.
We have started to use docker a bit and there are many projects in addition.
So the open source components we build upon enable what we like to call deployment agnostic cloud. Which means that it can suit the needs of the customer, the customer can select the hypervisor or storage technologies that fits them, the size of the implementation can be dynamic and can grow with the customer or can change in nature, so it can initially be deployed as a private cloud and then be gradually extended to a hybrid cloud model.
So for instance in our case we have made contributions to some of the open source projects that we build upon, and most specifically the CloudStack, where we helped build some of the AWS compatibility layer which is a key component in our product.
Now I want to spend some time to look at the crystal ball,
talk about the future of the cloud, and my predictions on how the cloud industry will unfold in the coming decade.
If we look back into history with the development of IT infrastructure we began with a totally centralized model with the mainframes of the 60s and minicomputers of the 70s. Then we went into the client server era of the 80s, 90s and up to the 2000s. The centralization trend started with the web in the 90s really and then there came this concept in the 2000s that was called the ASP or the Application Service Provider, if anybody remembers that terms, which kind of was the first form of SaaS.
So during this last decade we’ve been seeing strong trends towards a centralization of applications but I think the evidence is pointing towards that we are heading back towards decentralization.
With this I’m really talking about the cutting edge of cloud deployments, that largely will be used by early adopters like startups but others will trail behind this evolution by a decade or more.
If we look at the development of the cloud then really the precursor to the cloud was in house virtualization, some people like to call this cloud still today I don’t know why.
Then the phase came with the cloud adoption which we are in still today, we’re we mostly have the deployment options of the public and the private clouds.
The next step that will follow is the hybrid cloud which we are starting to see a little bit of, where you are starting to decentralize your application again.
And then the last phase is the development to a distributed cloud architecture that I like to call a cloud federation.
So this is my prediction, that the cloud is heading towards a radical new path which is towards decentralization, so that all cloud applications will become distributed systems.
And if you think about it then when you first saw somewone draw up a picture of a cloud on a blackboard then that usually meant a metaphor for the Internet.
And to add to that then the Internet is really the worlds biggest distributed system, so it makes sense that the cloud follows the Internet and becomes again really a truly distributed system.
We really need a term for this new cloud system, some call it cloud of clouds, or a cloud federation.
Some have used the term Intercloud, but that is used by some companies in marketing so I try to avoid that.
The key ingredient in the development towards hybrid and federated cloud is really identity management and identity system.
or on a broader scale identity federations, which are in my opinion the most important aspect in building up an ecosystem of hybrid clouds which will in turn end up in what we can talk of as cloud federations.
Because in the end it is all about user management and associated trust.
Some of the most popular identity providers used on the internet are the social medias but maybe some other players will come into the field to provide this role.
We have worked quite a bit with this specifically for a few customers in the education sector.
The education sector has built up a sophisticated federation system that spans across whole countries or whole continents, and I think this will spur unprecedented new collaboration and even sharing of resources such as on the cloud front, where you could start to see more of community compute clouds and shared resource pool between research organizations.
the Netherlands where we have integrated with the Dutch Surfconext federation, but the exact same methodology for the nordic research and academic community as each an every NREN has already established a federation, and I think even more important platform in this development will be cross national federations such as the Kalmar2 federation or the pan european edugain federation. And in my opinion this will propel a whole new wave of cross academic collaboration, especially on the cloud front.
In the future the clouds will be much more local, and many of these players will be smaller than the big players today, but can compete on different grounds than the economies of scale that the big players have such as locality, energy source, cost or some kind of resources.
When these players are acting more local you can much easier start to tap into local resources of each geographic location such as renewable energy, because there is renewable energy in abondance all around the world in forms such as hydroelectric, wind, solar, geothermal and so on. The picture here is actually from Krafla, which is a geothermal power plant in the North of Iceland.
Like you can see there are massive amounts of renewable energy widely available in the world, all the blue areas are areas that have more than 50% rewnewable energy available, but you can also notice that most of these locations with the blue color don’t have so much big cloud players active in these countries.
In our product we’ve built in monitoring of sustainability metrics that we use to present to our customers.
We display in our user interface metrics for carbon saving, carbon footprint and various other things that are calculated from the energy consumption used for the services.
We are have on the roadmap to release an open source version of this that can plug into the standard apache cloudstack project that we have integrated.
So in the future we will end up having is a distribution of the cloud ecosystem to be more reflective of where the actual population of the planet lives. And many of these players will act both globally and locally.
So what are the technologies that will propell this evolution.
I predict that some of these technologies will be instrumental to the coming trend of a distributed cloud architecture.
CoreOS and Docker are a perfect companion in implementing this kind of architecture, and even though these projects are really young they show tremendous promise in being able to deliver this in a very simple and elegant manner.
In addition to Docker there are really exciting project that show great promise such as Kubernetes, which was only launched in June, so it’s only about 4 months old. Mesos is also really interesting as it has some of the same goals and lines really well with a containerized and distributed solution like Docker.
I also suspect that the development happening around the distributed blockchain protocols like Bitcoin will play a role in this because a distributed cloud will need a distributed payment mechanism.
The big cloud IaaS players are racing to support these technologies, but I suspect that this next phase of the cloud will benefit much more the smaller players because it has the possibility of leveling the field when these become commodity technologies.
And of course all of these technologies are open source.
Google Borg
http://techcrunch.com/2014/08/28/microsoft-azure-now-supports-googles-kubernetes-for-managing-docker-containers/
What is really exciting about this all is that these developments somehow seem to be linked to Google. Google has of course been running a vast infrastructure on a global scale since its inception and has developed the technology and architecture for managing this with low cost, sometimes this datacenter operating system has been called Google Borg. And the CoreOS team has been honest about giving credit to Google for their inspirations. So things like etcd and kubernetes are both either inspired by Google or directly donated from Google.
So all of this will basically enable everybody and anybody to have google style technology at their fingertips.
So all you need in an internet connection, a few computers and these core components, and anybody can set up their own cloud infrastructure wherever in the world.
Easy access to these standard commodity software components will make it just as easy to set up a cloud infrastructure as a regular white box linux machine.
So what will happen is that we will se a vast global interconnected cloud of clouds that you will seamlessly be able to deploy your application, usually in the form of containers like Docker containers seamlessly on this gobal grid.
So to recap then basically all of the current cloud development has been built around open source projects, but I think in the future this will be even more pervasive and pretty much all of the core infrastructure software around this will be only pure open source projects and nothing else.
We couldn’t really have developed the QStack all on our own, the driver of this innovation was made possible through the variety of excellent open source project that we combine to build a better whole.