Thank you for coming today. We’re going to talk today about the next wave, Big Data Predictive Analytics. I’m going to talk about three problems where I think big data can play a significant role in helping us tape-out and build chips.
The first problem is basically Tape-Out Schedules. We generally find the problems, the challenges and the delays in our schedule a little bit too late to really do anything about them. 77 percent of the folks in our survey of 361 respondents said that tape-out schedule was the Number 1 thing that they would like to accurately forecast.
The way we do it now, it’s a very human intensive process. It’s a drag on development; engineers spend a lot of times in meetings and doing reporting, and its very error prone. The result that we get out of it, is basically that we find out a little bit too late to actually take any serious remedial actions.
The second problem where big data can really help us is Resource Allocation. We don’t necessarily know which design blocks are behind and which ones are on track. We don’t know which ones are in trouble and could use extra resources to keep them on schedule.
We’re trying to get all these blocks to cross the finish line at about the same time, in addition to the overall design, and we may need to reallocate resources to do that.
The data is actually in all of the log files. The data is actually in all of the data we have, but it needs to be mined and analyzed. Big data is a perfect application of this and we can help address this problem.
The third major problem where big data can be applied, and which we’re excited to be doing at IC Manage, is Insider Threat.
53 percent of folks surveyed in this US Secret Service Study acknowledged that they had had some type of internal data breach. We’ve all seen the Avago Skyworks Theft. There are many, many more thefts that go on in the semiconductor industry that are not disclosed. It’s a serious problem.
Big data analytics can analyze and detect when this is happening, and also prevent it from happening, with the implementation of simplistic rules that can be applied to stop IP theft. So three big problems where big data analytics can be used to make the semiconductor industry more efficient and more effective.