• Subscribe

Are you ready to achieve new levels of real-time decision intelligence in your factory?

New SmartFactory message bus – critical to enable advanced automated manufacturing

The SmartFactory Computer Integrated Manufacturing (CIM) solution allows manufacturers to define, control, automate, monitor, and record the entire semiconductor manufacturing process from front-end wafer fabrication through back-end assembly, test, and packaging. It does so through a portfolio of integrated software products that share information with each other via a common message bus that enables communication between CIM systems and applications.

At the heart of the CIM is the Manufacturing Execution System (MES), which is the master coordinator of processing throughout manufacturing. The MES is the key integration point with all other CIM systems for product as it moves through manufacturing, and the message bus is central to that integration, enabling communication between systems and applications. As manufacturers have moved from a mix of manual control and low levels of automated manufacturing toward higher levels (see figure 1), greater demand has been placed on the message bus to do more, faster. The existing messaging bus is limited in that it doesn’t provide the features required to achieve the real-time decision making necessary to advance automation to its highest levels.

Next generation message bus

To meet market demand and alleviate customer pain points with the existing message bus, we have developed a next generation solution built with distributed messaging architecture. The new solution addresses the shortcomings of the existing messaging system and allows for future expansion and technology modernization.
Figure 1: Technology Maturity – Manufacturing Automation Levels
Figure 1: Technology Maturity – Manufacturing Automation Levels

Pain Points

The message bus must enable real-time intelligence (managing real-time decision making and real-time data transfer) to achieve high and full levels of automated manufacturing. With the combination of real-time transfer of data and the ability to stream events, such as data from a tool or sensor, AI can analyze data, decide and take actions (as shown in figure 2). A message bus that is too slow in transferring data causes a delay that makes it no longer a real time view.
Figure 2: How the Message Bus Enables Real-time Decision Intelligence
Figure 2: How the Message Bus Enables Real-time Decision Intelligence

Scalability is also required to keep up with current market demands. There is so much data coming from each tool, every nanosecond, that there needs to be a very big pipeline for processing that data. The message bus needs to handle high message throughput, be scalable in the number of topics it can handle and accommodate rapid growth with low latency.

Customers also cite a need for multi-tenancy— low operating and maintenance costs, efficient use of resources and larger computing capacity. It’s also critical that a message bus operating at such a high level be durable, with guaranteed message delivery, zero data loss and failure recovery.

The other two pain points we sought to address were the need for geo-replication, the easy replication of message data between different regions and across different private or public clouds; and the need for a unified messaging and streaming platform for publishing and subscribing, storing and processing streams of data at scale and in real-time.

SmartFactory Message Bus Pulsar

In creating a next generation message bus solution, we leveraged Pulsar, an open-source software, and integrated it with our SmartFactory CIM through proprietary Pulsar adapters. We conducted a study two years ago comparing Pulsar to other similar types of message bus systems and Pulsar greatly outperformed the competition in the key areas we needed to address.

Our two-part solution of Pulsar and SmartFactory Message Bus Pulsar Adapter was designed with the capability to resolve customer pain points as well as meet future needs. It brings five major technological advances to our CIM messaging capabilities:

  • Multi-Tenancy: Pulsar was designed for deployment as a hosted service for private and public cloud with multi-tenancy architecture.
  • Scalability: Pulsar provides seamless scalability out to over a million topics with very low publish and end-to-end latency.
  • Durability: Pulsar uses a modern architecture (brokers/bookies) for optimal performance and resiliency. It provides guaranteed message delivery with persistent message storage.
  • Unified Messaging Model: Pulsar generalizes the two messaging concepts of queuing and publish-subscribe through one unified messaging API, enabling more use cases.
  • Geo-replication: Pulsar offers geo-replication as a first-class feature, allows
    organizations to deploy Pulsar across different cloud providers and replicate data across multi-cloud without locking in proprietary cloud provider APIs.

Conclusion

This robust modern message bus solution within a highly automated CIM solution will sustain future technology advancement inside the semiconductor factory. It will enable factory real-time decision intelligence and allow customers to improve process efficiency. Ultimately, the solution will empower digital transformation from a low level of automation to self-actuating automation.

About the Author

Picture of Bing Wang, Director of CIM Solution
Bing Wang, Director of CIM Solution
Bing is a seasoned professional with close to two decades of expertise in semiconductor manufacturing and fab automation software systems. Currently, she holds the position of Director of CIM Solution at Applied Materials Automation Products Group. Bing is at the forefront of driving innovation and interpreting industry trends to ensure that factory automation solutions remain cutting-edge. Prior to her role at Applied, Bing worked at Micron Technology and IM Flash Technologies, where she led software solutions teams and delivered groundbreaking automation solutions that set new industry benchmarks. Notably, she played a vital role in several successful fab startup projects for both Micron and Intel, which became significant highlights in her career.