Sanjeev Sharma (Terminus Circuits)
In recent time, 3rd party silicon IPs have proved to be cost-effective and catalyst for accelerated product development. Most of the program managers rely on well proven IPs with history of qualified customers and production data. The usual evaluation process is entirely skewed to a matured IP rather a newly developed IP.
The newly developed 3rd party IP comes with several risk factors, but it may also offer benefits like better performance, lower silicon area and better ROI proposition. A well-judged methodology to evaluate a 3rd party silicon IP may benefit from Time to Market perspective and it can reduce substantially reduce the NRE cost. Being an early adopter of this technology also helps the customer to work with 3rd party vendor for optimizing the IP as per their product requirement and hence avoiding the additional NRE towards customization.
Many customers have a detailed procedure for comparing and selecting 3rd Party IPs. Often this process run into time bound process and decisions are taken based on the “Performance” at the additional cost or “Low risk” at pricy point. Sometime customers just drop an idea of procuring the 3rd party IP due to non-availability of a pure rational method of selection criteria of 3rd Party IP.
This paper emphasizes the challenge and opportunities in the evaluation of new IPs being developed by 3rd Party vendor. Author suggests Perceptual Mapping technic for deciding the vendor offerings and project need. Also suggest some of the risk mitigations for a successful integration of the 3rd Party IP into the ASIC/SOC product.
This paper also recommend a practical engagement model from both engineering and business perspective for improvising the quality of IP deliverable and reducing the overall NRE cost of the IP development.
At the end, the paper discusses the adoption of complex IP, with the help of a business case in which the end customer takes a judicious decision and mitigates the risk at every level; hence successfully integrate a complex technology in his ASIC product. This case will also highlight the avenues for leveraging the development to reduce the NRE development cost and develop complex IP in most optimum way.
Perceptual Mapping has been several decades’ old technique to communicate the market competitions and criteria used by buyer to buy a product or services. It has been proved as a strategic tool for identifying the differentiator by product companies or a service company. Semiconductor IPs is always a tough choice for many program managers. There are many obvious and non-obvious factors of a Semiconductor IP that contribute to the success of any project.
Often buying decisions for 3rd Party IP is largely dependent on the profile of the IP vendor and its reputation rather than the intangible or tangible benefits of the new IP product.
Also, since the buying decision or criteria of one company differs from the other company. It is difficult to generalize the evaluation criteria of 3rd Party silicon IP. There are several risks a program manager can foresee.
Figure 1 Sample Risk Evaluation Sheet
How perceptual mapping helps the customer in buying decision
Based on the engineering team preference, the program manager can either fill his evaluation sheet for IP with respect to various vendor options on a merit based grading or he can do a strategic analysis of perceptual mapping.
Below Figure 2 illustrates a perceptual mapping for a buying scenario of a 3rd Party IP. Though the Euclidean distance for Vendor A is smaller than vendor C, but it is worthy for a buyer to negotiate with Vendor C to improvise the performance of IP and mitigate the risks factor than settle with vendor A and buy a pricy IP (size of bubble denotes the pricing)
Figure 2 Sample Perceptual Mapping for SIP Product
How to mitigate the risks and improvising the scope for better performance
As suggested above, until the Buyer and IP vendor work together to reduce the risk and improvise the performance parameters, it may impact the project badly. Below recommendation may be useful in reducing some of the standard risks an IP buyer can foresee and improvising the performance of IP.
- Reference check about the IP vendors and their key team member
- Financial balance sheet study
- Vendor’s end customers market success
- Analyses cost differentiation carefully. It should not come at the cost of quality
- Analyses vendor’s partners and contractor’s profiles
- Bank guarantee for the project
- Review the IP documentation, integration guideline, EDA flow document, system models, test bench, test plans etc.
- Project reviews at important milestone
- Review the design methodology and suggest if any changes is required
- Deliverable review with internal and external experts
- Review of pre-silicon verification and silicon validation methodology
- In case of significant customization, negotiate vendor for a pilot test-chip
The above list is an indicative list this may also include several other factors which might have influenced the project risk. Carefully prioritizing the risks and mitigate them at earliest helps the perceptual positioning of IP vendor on the perceptual map and hence it can come closer to the ideal one.
There are several factors may be hidden or resurface once the source IP has bugs or there are integration issues. This may cost several months of product re-spin and millions of dollars of additional expenditure if the risks were being ignored. Many vendors avoid additional customization or customers request to accommodate the additional design/verification flow. This indeed an additional NRE cost for Vendor company. Negotiating wisely on the business model may also help the program manager to optimize the IP cost and re-map the vendor positioning on high performance vertex.
Interface IPs are important silicon content of most of the networking and communication IC products. In these products, the data rate or throughput mainly depends on the merit of interface IPs. This case illustrates a collaborative approach for a development and deployment of a new complex IP in one of the communication product of Terminus circuits’s partner company. One of the prospective customers of Terminus Circuits approached us for the requirement of the PCIe 3.0 PHY IP for their communication product. Terminus Circuits has previously developed similar product in other foundry node and the IP needs to be redesigned/ported into the newer process nodes. The partner company did a due-diligence and identified key risks in the engineering engagements. Since the IP had to be ported into a specific process node and it was also lacking the silicon history, the overall risk in the IP purchasing was quite high. The company shared its risk assessments to Terminus circuits and seeks Terminus Circuits opinion in mitigating risks. Considering the team profile and competencies of Terminus Circuits, the company decided to work with Terminus circuits and agreed for a joint development. Below model illustrates the mitigating plan for minimizing the risks and improvising the quality of the IP.
Figure 3 Engagement model for Mitigating Risks
How Perceptual Map changes during the IP negotiation and execution
After negotiation and pre-finalization phase, it is important to re-visit the perceptual map in order to understand about the new positioning of the vendors. In below map we can notice the new positions for both vendor B and vendor C. A thorough cost benefit analysis will help program manager to understand the real cost of IP procurement which includes the additional NRE towards resource bandwidth, yield, design success and other hold-up risks.
Figure 4 Perceptual Map after IP negotiation
Customers often defer the project due to non-availability of 3rd Party IPs at an ideal performance and price point. The above paper highlights the challenges in the evaluation of a newly developed IP and also suggests some practical way to do a strategic analysis for buying newly developed 3rd party IP. The author has also shared a case study of a complex IP business and its deployment. The bottom-line of this paper is to bring a perspective of a buying model for a risky IP and mitigating the risks by adopting some practical approach. Readers can use Microsoft Excel for producing Perceptual maps or use statistical tools for multi-dimension analysis.