MDBoyd

The end of the Pelco era

In 2013/2014 I was part of a research effort at Pelco that identified the drivers of new surveillance technology. We wrote two different white papers during that time, with me as the author and editor.

The research lead, Sam Grigorian, demonstrated the importance of asynchronous cores in video processing. Co-researcher Fida Almasri identified the necessity of accelerating Pelco’s production cycle in order to stay competitive, and I argued that cell phone processors could allow us to leapfrog new surveillance systems and be our new platform for advanced surveillance hardware.

Our white papers explained that we are at a very interesting and challenging place in Moore’s Law.

In tech forums I often see people write about the end of Moore’s Law as if it meant the end of innovation. I often read people who claim that the limit of miniaturization is coming “soon”.

The truth is that Moore’s law is slowing down. Processing cores have just barely made it to the 7 nm node of semiconductor device fabrication. 5 nm semiconductors have been constructed and demonstrated in the lab. We may see these in commercial production around 2020.

The Lawrence Berkeley National Lab demonstrated last year that it is possible to create a transistor with a working 1-nanometer gate.

If researchers solve the problems of quantum tunneling and of removing heat from processor cores then we could see miniaturization of up to 3 to 1 nm before we hit that final wall of physics in this technology.

And then we will need to advance in other technological directions to allow for the increase of switching devices per volume.

However, the end of shrinking transistors doesn’t mean the end of innovation.

Quite the opposite. As we pointed out in our white paper, technological innovation is accelerating. It is being driven primarily by cheap and easy to use processing cores.  The rate of innovative change has accelerated to a point where is is difficult or impossible to predict new technology further than 18-months out. And the rate of technological innovation seems to be increasing.

It has become much easier to solve advanced technological problems by throwing a handful of transistor dense processor cores at it.  But this leads to a different problem.  Multiple processors work best with parallel processing.  And our technology is still in the infancy of parallel programming.

Development environments for parallel programming are sorely lacking. Parallel programming environments are not easy or intuitive for human programmers. We are still coming to terms with the problems identified by Amdahl’s law – a formula used to describe parallel computing workload and used to indicate the hard limits on how many parallel processors may be used to solve a given processing task.

Even the latest and best Integrated Development Environments designed for parallel design require software “tweaking” and other kludges by the software engineer.  As the development environment matures we should see even more gains in innovation.

Back to Pelco

These were the challenges for Pelco in 2013.  Accelerated innovation necessitated a decreased production cycle.  Advanced surveillance systems required embracing asynchronous processing cores.  And video over IP was transitioning to advanced networks using better hardware and compression methods.

Pelco once dominated the security marketplace using advanced technology in a market that was traditionally low-tech.  So it is somewhat ironic that Pelco couldn’t or wouldn’t keep up with technological advancement in the security industry.

Faced with the urgency to dramatically upgrade their technology and production processes to keep up in this accelerating environment of innovation – they blinked.

And Pelco’s parent company at that point decided to let Pelco’s era come to an end.

At one time Pelco employed over 2,000 people.  There are under 1,000 people at Pelco now.  Another 200 will be laid off from Pelco during 2017.  And I’ve been told that the Clovis manufacturing facilities will be closed at the end of 2017.  Any employees not working for Schneider Electric – Pelco’s owner – face an uncertain future.  There are still a few product lines that are or will be transferred to other facilities.

The white papers that I helped create may have been one of the last straws in this decision.