Trend toward mass collaboration

petri Thoughts

Everyone knows the exploded growth in the game and software business in general. Especially gaming industry is hyped so that even the education is skewed towards games and software development. Who would have guessed that Universities of applied sciences educate engineers for Gaming industry.

What enabled the explosion of software business? Some of the biggest contributors are Open Source communities and their excellent development tools for software application business. Without Open Source development tools, it would be difficult for small teams to create games or other applications.

In other words, the best opportunities to add value to the industry happens, when there are equal chances and rights for everyone to be creative – like we have seen in the gaming industry. Public school in Finland is a good example of giving the chance to be creative and add value to the society. It is the crucial enabler and base for the mass collaboration.

It is clear that the application development is already well saturated where each day 1000 new applications are made. It is challenging to stand out from the crowd.

At the same time hardware based development is growing like Arduino based platforms are clearly showing. Already it is possible to find Open Source data for almost all aspects of electronic devices. For instance, there are Makers communities developing devices that are almost fully Open Source including plastic cover, PCB and software stacks. Basically ICs and discrete components are only off-the-shelf components that cannot be customized.

We could say that prosumers, which are the new generation of people that are consumer and producer  at the same time, are slowly conquering the whole chain of device development. Look how young generation is interacting with anything they touch. More and more, people want to be prosumers.

How about microchip business, such as ASIC design development in the semiconductor industry? During the golden hardware decades of 1990 and 2000 in Finland due to the Nokia, we saw the similar trend in hardware business where schools “manufactured” workers for Nokia purposes with engineer titles such as Telecommunication Engineer with major subject as ASIC development. The difference was that graduate students had to apply jobs for big companies whereas for gaming or web hosting business graduates are able to establish a company with minimal investment.

The whole semiconductor industry is very closed industry where each company protects their IPs better than gold in the bank. Also ASIC business is very expensive, the process to create a chip is fairly complex and time consuming. Licenses for commercial development tools are expensive. Flow environments are scattered and utilize multiple EDA tools. The flow environments are usually company specific and they require heavy scripting and resources to maintain. There isn’t simple and complete IDE for IC development to build chips rapidly.

At the same time ASIC designs are getting more and more complex. And it requires more and more resources to complete the development of a chip in reasonable time. Yet there isn’t enough resources to fulfill all wild ideas to chips.

It’s clear that it is impossible for an enthusiast or a small team to implement idea to a chip themself. However, there are parts in the ASIC process where a small team or a community could already contribute, such as IP design creation. So, there is a glimpse of hope for opening the industry for masses.

Let’s imagine that all these obstacles would be solved somehow and a community would be able to develop complete IP designs for ASIC business. It could be a tremendous boost for the industry where the design houses could outsource parts of the design blocks for masses. Already we have sites such as OpenCores.org that provide Open Source IPs available for community or design houses. Also there are some Open Source SoC projects that integrate Open Source IPs for FPGA. We could have IP libraries of lower level design components such as interfaces or processor cores as Open Source blocks to build bigger System on Chips. These free libraries would be then basic utilities to build bigger systems.

To speed the design cycle time, we could have Open Source based consistent and reusable environment or framework to orchestrate the IC flows to automate the chip creation. Even better, if IPs are hardened on certain technology nodes it would enable bigger SoC designs to be build faster. This could free resources for system verification, which is already the biggest contributor for the development cycle.

In this picture, communities or design houses could start a new project much faster by taking existing flow environment as their starting point. Then thye could focus more on the fast design integration without the hassle of flow environments. At the end, all this could enable chip development for everyone. Just like in the software application business.

So, what should be done to make this imaginary future vision more viable?

To change the trend from software business towards hardware, we need to have Open Source based tools, kits and libraries for users to develop ASIC design. To establish collaborative way of working, the industry requires good Open Source working environment that enables de facto deliverable exchange and orchestration of different flows.

Of course someone could argue that is it really necessary to open the IC development for everyone. Why not to keep the mystery of chip development for professionals?

Short answer would be to look on other industry fields that are already heavily utilizing mass collaboration to boost their R&D capabilities.

The future will tell if we are gradually seeing the change towards mass collaboration for the IC industry.