IBM has unveiled the latest addition to its Z series mainframe-class computer, the z16, which the company has positioned as a powerhouse designed to combat fraud.
Ross Mauri, general manager of IBM Z, said the on-chip artificial intelligence (AI) accelerations support up to 30 billion deep learning inference requests a day at 1 millisecond latency. “There are around 8 billion in the world,” he said. “If everyone could go on a shopping spree simultaneously, we could handle it.”
IBM said the AI inferencing, via its Telum processor, combined with high-volume transaction processing, enables banks to analyse fraud during transactions on a massive scale. “For consumers, this could mean reducing the time and energy required to handle fraudulent transactions on their credit card,” it said. “For both merchants and card issuers, this could mean a reduction in revenue loss as consumers could avoid frustration associated with false declines where they might turn to other cards for future transactions.”
In a blog post discussing the z16, Mauri said IBM had updated Watson Machine Learning for IBM z/OS to help clients make faster predictive decisions using insights from operational data, including native z/OS applications. “This is critical for use cases like fraud detection, where bringing AI inferencing to the transaction before it completes would help businesses analyse for fraud on a massive scale,” he wrote.
IBM has also launched DB/2 13 for z/OS, the latest release of its enterprise database.The new database system uses the Telum processor to power new functionality called SQL Data Insights, which Mauri said helps to identify similarities and clusters in data.
IBM distinguished engineer Elpida Tzortzatos said businesses can derive value from AI by accelerating fraud detection. “Instant payment requires a real-time machine due to increased fraud,” she said.
This, said Tzortzatos, is what IBM has been able to deliver, thanks to the z16’s on-chip AI acceleration engine, which has direct access to the computer’s level 2 cache. This enables IBM to reduce overheads normally associated with transferring data from processor memory to a GPU’s (graphics processing unit’s) memory to run AI algorithms, she said.
Ric Lewis, senior vice-president at IBM Systems, said: “IBM is the gold standard for highly secured transaction processing. Now with IBM z16 innovations, our clients can increase decision velocity with inferencing right where their mission-critical data lives. This opens up tremendous opportunities to change the game in their respective industries, so they will be positioned to deliver better customer experiences and more powerful business outcomes.
IBM claimed that z16 is the industry’s first quantum-safe system and has been engineered to protect data against future threats that could evolve with advances in quantum computing.
It uses what IBM describes as lattice-based cryptography, an approach for constructing security primitives that helps protect data and systems against current and future threats.
IBM said the new hardware provides secure boot, which when combined with quantum-safe cryptography can help businesses tackle threats such as harvest now, decrypt later attacks, which lead to extortion, loss of intellectual property and disclosure of other sensitive data.