Frequently Asked Questions
What are you selling?
Basically, we're selling a meme -- a powerful, contagious idea.
This idea is that networked computing doesn't have to be done the
way we do it at the moment. And that there are good reasons --
technical, social, environmental, economic -- for doing it
differently. In concrete terms, we believe that ultra-thin client
computing and Open Source software can deliver affordable and
robust networked computer workstations, of benefit throughout the
world but particularly in developing countries. See also our
Project Intro.
What is an Ndiyo system?
When we refer to an Ndiyo system, we are talking about a server PC, running Open Source software, driving a cluster of ultra-thin-client workstations. We've developed a particular solution, using the Ubuntu Linux operating system and the new Nivo hardware that we have developed with our partner DisplayLink, that provides a solution that is cheaper than any existing way of providing networked computer workstations, while providing good performance and high usability for typical office or communications use. See also our page on the Ndiyo system.
Is the Ndiyo project tied to a particular vendor?
No. The Ndiyo project is a non-profit project, fundamentally independent of any commercial interests. But we felt that existing thin-client hardware didn't live up to our vision and we were conscious that nobody would take our idea of ultra-thin-client computing seriously unless we could demonstrate that that it really works, technically and economically. So we decided to create a company to design and build kit that would enable our vision of an architecture that is spare, elegant and economical. Thin-client networking can be done in lots of different ways -- for example by using kit from Axel, DisklessWorkstations/LTSP, Sun, HP or IBM or by re-using old PCs. At present we believe that the Nivo technology that has emerged from our collaboration with DisplayLink will offer the best price/performance ratio, but we are also evaluating hardware from other sources. We're pretty demanding, though - it must be both very high performance and very low cost.
DisplayLink, however, just produces the underlying chip technology. More recently, we've created Camvine, a company which is taking that technology and manufacturing it in an Ndiyo-compatible format for use in other products. While Camvine is an independent for-profit company, the founders arranged for Ndiyo to own shares in Camvine, which means that if Camvine is a success, Ndiyo will also benefit in the longer term.
How is Ndiyo funded in the meantime?
The project has been funded to date by philanthropic donations from a number of people who have personally benefited from their inventions and investments in technology. We welcome offers of financial and other support. If you are interested in helping, please get in touch.
Where is Ndiyo located?
In Cambridge UK, though we have active supporters and participants in a number of other countries, including the US and Ireland.
What's your business model?
We're currently a non-profit organisation and may eventually become a charitable foundation, so 'business model' is not perhaps the most appropriate term. The nearest we come to it is a business metaphor based on a very basic industry -- plumbing. Nobody owns the intellectual property implicit in plumbing by circulating water through pipes, but all over the world people make good livings from it: some make taps, pipes and equipment; some profit from retailing plumbing equipment; and plumbers make money by providing services to consumers. We are aiming to achieve something analogous -- to create an effective IT networking architecture and make it available to the world so that other people can profit from exploiting our ideas without being able to exclude anyone else who wishes to use them.
We do believe, however, that for initiatives such as ours to be effective and scalable in the longer term, they need to be economically self-sustaining. So we're always thinking about ways in which Ndiyo might become a viable business without losing its humanitarian agenda.
Isn't the thin-client idea old hat?
It's a very old idea, for the simple reason that it makes sense. It's had a bad press because earlier attempts to create working thin-client networks were over-hyped and tended to give disappointing performance. Fundamentally, we believe that to achieve our vision of a truly affordable thin-client terminal we had to design a piece of hardware using a completely different design philosophy.
The nivo client that has emerged from our collaboration with DisplayLink has a very different design philosophy to existing thin-client hardware. Whereas most designers started with a PC and then looked for components they could take out, we started with a bare monitor and asked what we would need to add to create a working thin-client terminal. The nivo is a specialised electronic unit carefully optimised to the task of pulling pixels over a network!
What makes you think Ndiyo can do better?
Because we're doing it right now. We kept quiet about the project until we were sure that Ndiyo-type thin clients could deliver a user experience comparable to that obtainable with conventional technology. We didn't want to be accused of promising well-intentioned 'vapourware'. Now Ndiyo systems are being installed in real-world environments. And they work just fine.
Who else is doing thin clients nowadays?
Lots of people, including big players like Sun Microsystems, HP, Wyse and IBM. These companies are producing thin-client solutions because the technology appeals to large organisations who are finding that the burden of administering, supporting and upgrading systems with thousands of networked PC is becoming prohibitive. But we're interested in much smaller organisations -- companies employing less than 50 people, schools, universities, Internet cafes and households -- which need networked systems but lack the IT skills and budgets needed to deploy conventional technology.
What happens if my Ndiyo server isn't powerful to support all its clients?
Two options: (a) buy a more powerful machine; or (b) put another PC alongside it to form a Linux cluster. We're working on a system to make clustering simple to set up (we call it 'plug-and-play clustering'). The point is that, unlike conventional PC-based networks, if the system becomes too slow you don't have to replace every thin client.
Are you trying to replace the PC?
No. What we are seeking to replace is the wasteful and cumbersome networking architecture based on linking PCs together. A better way of looking at an Ndiyo system is as a 'PC multiplier'. With the right operating system, most modern PCs are powerful enough to support five or ten users, but in general each machine is used by only one person. With the Ndiyo approach, almost any PC can serve several users -- and provide them with the same level of service they get from a standalone machine.
Why don't Ndiyo systems use wireless networking?
We have systems running on wireless networks, but the bandwidth isn't yet sufficient to give a really seamless user experience on the clients. In applications where there isn't much interactivity -- i.e. relatively little data coming back from the client to the server -- that doesn't matter. But for office applications at present the performance over wireless isn't quite good enough, and we don't want Ndiyo users to sit there wishing that they had a 'real' PC.
Wireless bandwidth is improving fast, though, as is the ability of DisplayLink's underlying technology to take advantage of it, and we expect wireless Ndiyo systems to be an option in the near future.
Will your system work with Windows?
Yes, technically there's no problem. But from the point of view of the Ndiyo project, running Windows on our systems militates against our desire to be free of proprietary software, and also increases the costs dramatically because of the licence fees involved. Each client, for example, needs a Windows licence, the server software is licensed on a per-seat basis, and multiple licences for Microsoft applications software dramatically push up the cost -- and reduce the scalability -- of a thin-client system.
The majority of the world's Windows users, of course, are running pirate copies for which they have not paid, and this does help reduce costs! But we don't feel it's a good model to be promoting for the world when legal, and in many ways superior, alternatives exist.
Why not just use reconditioned or discarded PCs as thin clients?
This is an approach extensively used in developing countries at present as industrialised countries unload their obsolete PCs ahead of the implementation of the EU Waste Recycling Directive. But our discussions with government and NGOs tell us that using refurbished PCs involves a lot of maintenance, support and general hassle. A typical PC also uses about 20 times as much electricity as a Nivo, which can be a significant cost that is often forgotten.
What kinds of applications do you envisage for Ndiyo technology?
We started off with ideas for three kinds of packaged networked systems:
- A 'classroom in a box'
- An 'Internet cafe in a box'
- An 'office in a box'
These were intended to demonstrate how compact an Ndiyo network can be -- you could fit an entire five-screen office system in a large cardboard box about the size of those in which large plasma TVs are delivered. Our friends in the NGO world have suggested another application -- a 'Humanitarian Information System' (HIC) in a box. HICs are what NGOs and disaster relief organisations set up on location. They are used, for example, as a way of building databases of missing or displaced persons so that families can be reunited. Setting up an HIC with conventional PC-based networking in disaster zones is logistically and practically difficult. Our more compact systems may well prove useful in such settings.
Does an Ndiyo system have to have flat panel displays for the clients?
Absolutely not. If a device can display pixels and has a VGA cable, it can function as an Ndiyo client. That means, for example, that you can build an Ndiyo network using CRTs that companies are throwing out.
Can the nivo ultra-thin client be powered using Power over Ethernet?
We could power it that way, certainly, but since there needs to be a mains power supply for the monitor anyway, we don't think there's sufficient benefit to warrant the substantial extra cost.
What about support?
We're designing the system in such a way that the servers can be remotely administered, if desired, and are exploring financial models for making that possible.
When will the Ndiyo hardware be available? How can I get an Ndiyo system?
Our current system uses remote-display technology developed by DisplayLink, a company which grew out of Ndiyo. Prototype Ndiyo hardware is manufactured and sold by Camvine, another Ndiyo spin-out.
Please see the Availability page for more info.
Is there a relationship between Ndiyo and the $100 laptop Project?
The two projects are entirely separate. They have a common general aim -- to tackle some aspects of the so-called 'digital divide', but the One Laptop Per Child (OLPC) approach focusses on providing mobile devices for children, whereas Ndiyo is focussed on rethinking networked computing to make it:
- affordable
- supportable and
- environmentally sustainable