I will show you an interesting experience about cloud computing market. If you try to order a cloud server from Alibaba in the USA, it will cost between 4.5$ and 79$ per month.
And in HK, it will cost between 9$ and 79$. That is twice as much.
In China it will cost between 19$ and 109$. That is four times more.
Let us now have a look at what we can get in Europe from a French company called Kimsufi.
For less than 20$ per month, we get a server which is two times bigger than the one Alibaba provides in China. And there is no limit for network transit.
Let us now look at Rapid.Space, a cloud provider which my company is operating.
For about twice as much as the largest server of Alibaba in the previous screenshots, we get 10 to 40 times more performance. Again, there is no limit for network transit, unlike Alibaba cloud.
Each time we compare prices of cloud in Europe and in China, we reach a simple conclusion: China cloud costs 5 to 20 times more than European cloud.
Could we change that?
Some people say that price of cloud is not important.
This is not true. Excessive price can prevent economic development.
Imagine if the price of cars in China was 10 times higher than in Europe. Everyone would just ride bicycles. There would be no highways.
In Europe, cloud is so cheap that every student can afford his or her own cloud server, then learn by experience. This is close to impossible China. As a consequence, European students have much higher skills about cloud, network and cloud software than students in China.
Besides students, if price of cloud in China could be lower, more than one million jobs could be created, as we calculated in an article published on our web site..
In our presentation, we will first introduce Nexedi, our company.
Then we will introduce our expertise about cloud and edge computing. Nexedi created the first edge computing software called SlapOS. SlapOS is one of the very cloud management software which really works. We were recently invited in Hangzhou to introduce SlapOS at a conference co-organised by the Chinese academy of science.
We will then introduce some suces cases of SlapOS.
At the end we will introduce yr plan for Rapid.Space China, a company that was recently created in Shanxi region and which we are trying to support.
Nexedi is probably the largest Free Software publisher with more than 10 products and 15 million lines of code. Nexedi does not depend on any investor and is a profitable company since day 1.
Nexedi has subsidiaries in Japan, Germany and China which all focus on entreprise services based on Free Software. We have a subsidiary called VIFIB which provides cloud management services, such as those used for Rapid.Space China. And we have a minority share in ClearRoad, the first company in the world to provide a clearinghouse for road usage charging in the USA, a new way to manage road congestion by making drivers pay for creating congestion.
Nexedi clients are mainly large companies and governments looking for scalable enterprise solutions such as ERP, CRM, DMS, data lake, big data cloud, etc. Nexedi clients include Airbus, SANEF, Mitsubishi, Kyorin, PSA, Sankei Chemical, etc.
Nexedi has more than 10 years of expertise in cloud and edge computing.
SlapOS is the first edge computing software in the world. It is a Free Software developed by Nexedi that can be used to manage edge computing system, cloud computing, big data centers or orchestration services. If someone needs to create a new cloud service, SlapOS is the best way to start because it covers all aspects, from order taking to billing through provisioning and billing.
SlapOS first appeared in the media in 2009.
Nexedi has been using SlapOS since around 2008. SlapOS supports many kinds of applications: CDN, database, virtualisation, big data cloud, ERP5, IoT, etc.
We call it "Edge Computing" because SlapOS does more than cloud computing (see Five Evolutions of Cloud Computing "https://www.nexedi.com/NXD-Blog.Five.Cloud.Evolution").
Unlike most cloud computing systems, SlapOS can manage cloud resources not only in single data centre but also in thousands of small data centres located in different cities, regions, countries, etc.
SlapOS: a Multi-purpose Distributed Cloud Operating System Based on an ERP Billing Model
SlapOS also appeared in scientific publications.
SlapOS ideas originate from a discussion with Professor Christophe Cérin - who is currently visting professor at HDU (Hangzhou) - about grid computing systems used by NASA for example. Cloud was very popular at that time (2009) and Nexedi was already using it.
It became obvious during this dicussion that the main purpose of cloud was to automate things, rather than virtualise things. It also became obvious that nothing prevented from doing cloud in a decentralised or federated way. IPv6 had become available in all homes in France. It was therefore possible to host servers at home and build a distributed cloud, or to distribute cloud service in many datacenters.
SlapOS has already been taught in universities.
The first lecture about SlapOS was given at HDU with Prof. Congfeng Jiang in Hangzhou.
Some of the students we trained now work at Qiniu or large IT companies in China.
SlapOS was designed to be resilient.
It is capable of being operated in very harsh conditions:
We designed SlapOS like this was because we found that many cloud providers are not reliable. The article "Downtime statistics of current cloud solutions" (http://iwgcr.org/wp-content/uploads/2013/06/IWGCR-Paris.Ranking-003.2-en.pdf) gives a good overview of the problems that most cloud providers face in terms of networking and electricity no matter, who is the supplier. All cloud suppliers in the world face downtime of a couple of hours per year.
We tried to solve such problems with SlapOS. We also solved the problem of unstable networking which every professional faces in China due to congestion or routing errors.
SlapOS was recently introduced to the German industry at the Edge Computing Consortium Europe in Berlin. It could become in the future a key component for open source edge computing architectures used in Industry 4.0.
What makes SlapOS so different is its simplicity and resiliency.
Its simplicity is the result of using only two components: a master and a slave. Other system can have hundred components.
For resiliency, we based all our design on the idea that resiliency must be implemented with software and should rely on a distributed infrastructure of distant sites with diverse suppliers of telecom or electricity. However each site or hardware does not need to be redundant. This is the same as having multiple imperfect suppliers rather than expecting a single supplier to be always perfect.
By sticking to a very simple and minimal architecture, SlapOS could also achieve with a small budget what huge community projects such as OpenStack still fail to achieve after 10 years. And we could do much more, because our architecture was more generic.
Thanks to simplicity and resiliency, SlapOS system is much more reliable than competing technologies called OpenStack or Docker, and provides much more performance.
We did some statistics on hundreds of servers, some of which run SlapOS and others run competing technologies. The result is that competitors face instabilities between 3 and 10 times more frequent than with SlapOS.
The reasons are explained in an article: "Five Evolutions of Cloud Computing".
Another advantage of simplicity is that SlapOS is also more universal. It can be deployed smartphones, 5G base stations, big servers, drones, satellites, etc. using the same approach. This opens applications far beyond traditional cloud. It can be applied for example to IoT, industrial automation, smart cities, aerospace, etc.
SlapOS was actually deployed some years ago inside a 777 flight of JAL between Paris and Tokyo, serving real time web content during the journey.
Let us have a look at some success cases that are based on SlapOS.
SlapOS is used by Teralab. Teralab is the big data infrastructure created by the French government for data science and AI. See: https://www.nexedi.com/success/slapos-IMT-Documents.Teralab.Success.Case.
SlapOS is used by Rapid.Space
Rapid.Space (https://rapid.space/) is a high performance, low cost cloud infrastructure that provides:
It is available in Europe and we hope soon in China. It is based on SlapOS and Open Compute Project (OCP) hardware.
Everything about Rapid.Space is public.
Grandenet is our partner in China. Grandenet provides acclerated IPv6 networking in China with accelerated Web sites. Grandenet has official government licenses. Grandenet is based on SlapOS.
Woefel (Germany) collects data from hundreds of wind turbines using fluentd which is itself deployed using SlapOS (server side).
We will now provide a brief overview of Rapid.Space China. Rapid.Space China is an independent company which goal is to provide in China the same kind of service as Rapid.Space in Europe, at similar price.
In order to make Rapid.Space China, we need to understand how to operate a low cost cloud provider.
The secrets of a low cost cloud are well known in Europe where cloud price is 5 to 20 times lower than USA or China.
It requires a site with low enough temperature, so that no cooling is required. Recent cloud servers made by the "open compute project" (OCP) are compatible with this requirement. The Shanxi province in China is ideal from this point of view since temperature is always below 35 ℃.
Cheap electricity is needed. Cheap land. Shanxi is again perfect for this.
Cheap hardware is required, which oftens means using recertified hardware from the circular economy. This can be challenging in China, but it should be possible.
Management software is essential and often forgotten. This is where Nexedi and VIFIB are very strong. SlapOS is based on ERP5 and thus supports all the billing system, an essential component for a cloud operator.
A simple offer is required in order to be understandable and cost less in support. A simple offer also means simpler management software, less costs.
There should be no redundancy to cut costs. If resiliency is needed, multiple sites should be opened.
Last, we need cheap Internet. This is impossible in China for the time being. So, we have to charge for transit.
Here is our plan.
In the first step, we will send a rack of servers in Shanxi and run some benchmarks.
We will then send a second rack in another city, for resiliency. The service can then be launched commercially. If one site in one city fails, a second site is still available.
We can then build a datacenter to host hundreds of racks, using a free flow architecture which minimises costs, electricity and environmental impact.
In parallel we sent racks to more cities in China, wherever we can get good price.
Last, we can build more datacenters, in China and abroad.
Investment for step 1 and step 2 have been made. We are now looking for step 3 for which we need 20 million RMB.
Through this presentation, we have demonstrated how we can change the current situation of cloud price in China. Thanks to the unique advantages of Shanxi region, it will soon be possible to benefit in China from the same competitive advantages of European cloud technology.
You can find more articles related to SlapOS: