Bulky mask data sets create storage problems

By Rachel Robinson
WaferNews Associate Editor

As mask data sets become more complex and data-intensive, maskmakers face the challenge of how to manage large amounts of data that must be moved from design to mask.

Tom Grebinski, chair of SEMI’s design/photomask data path task force explained: “The density of integrated circuitry has increased according to Moore’s Law, and subsequently, the amount of data required to express, and ideally resolve, all of the features of an IC design has increased as well.”

With this increased data comes three major focus points, according to Shawn Knox, VP of information systems at Photronics, a sub-wavelength reticle solutions supplier:

*How to move the data from the design centers
*How to process the information or put it into the proper format
*How to store and archive the data

“The amount of data needed to design and manufacture photomasks and ICs will continue to increase with time,” noted Grebinski. “The problem needs to be dealt with concurrently at the design level, the design data encapsulation level, the data preparation level, and at the computational data path level for both photomask pattern generation and inspection technology.”

Photronics is expecting one file of source data to swell to 324 gigabytes by 2005, more than three times the memory needed to store the entire Encyclopedia Britannica.

“Once the design data reaches Photronics, it must do data fracturing and prepare for inspection tools, and then for metrology tools,” Knox told WaferNews. “A critical layer node on file can be exploited three or four times.”

Grebinski further explained the problem. “The necessary addition of resolution enhancement technology accelerates the increase in data volume and thus the need for a substantially greater computational infrastructure.”

In hopes of addressing the problem, Photronics built a new wide area network (WAN) to link Photronics with its customers, based on broadband and Internet protocols.

Photronics’ Advanced Technology Data Center was developed to address increased data management complexities including IT infrastructure, data communications, design data transformation, storage and archiving data, redundancy and fault tolerance, frontend reticle data preparation, centralization, standardization, automation, and optimization.

Knox said that Photronics has built the WAN-based on existing communications infrastructure. “What’s important about the solution is that we have no single point of failure in the infrastructure. If we can’t move data, we can’t move product.”

Moving the data will prove to be a major problem down the line. According to Knox, the challenge in coming years will be how much bandwidth is available. Knox continued that infrastructure is constantly being built, and Photronics “has its eyes and ears on the telecom industry.”

In terms of handling the information once it is received, Photronics has invested in data storage equipment that allows it to store multiple terabytes of data. According to Knox, the data is stored in a central form on a regional basis. Photronics is using storage-area networking and network-attached storage.

“If you look at how much storage is required to push data to manufacturing tools, we’re talking tens and hundreds of terabytes,” he said.

“Seeing how traditional data movement is pushing Moore’s Law, it can become a constraint,” Knox cautioned. “Data management and handling is going to be a roadblock.”

From Grebinski’s point of view, the many problems associated with mask data set size and the movement of data is an industrywide problem. He warned that the majority of these problem sets are outside the scope of any one company or industry, but have everything to do with data inefficiencies and increases that are being faced today and will be faced in the future.

With an eye toward the future, Photronics is focused on a three-year business plan with flexibility being the key. “[We need to have the] ability to respond and meet the needs of business,” stressed Knox.


Easily post a comment below using your Linkedin, Twitter, Google or Facebook account. Comments won't automatically be posted to your social media accounts unless you select to share.