The use of micro-services architectures over cloud computing has become the industry standard for operating web and mobile applications as well as internal company computer systems. AI services are becoming a dominant factor, affecting cost/performance dramatically.

The Problem: Uncontrollable operational cost/performance

• The micro-service architecture “pretentiously” offers a panacea of performance/cost worry-free environment.
• In practice, system operators face micro-service coordination challenges resulting with uncontrollable operational costs or, alternatively with significant performance degradation. These are caused by i) Traffic fluctuations, ii) Miss-design, or iii) DoS attacks exploiting these weak points.
• System size (10s-1000s micro-services) complicates operation & coordination.
• AI: The current trend of incorporating AI (LLM) micro-services into any system, is likely to increase the associated costs and the problem criticality by an order of magnitude.

The Solution: A holistic dynamic resource allocation technology  

• We propose a unique dynamic and holistic approach for effective cost control with little effect on performance and with attack resilience. It is based on two main synergetic building blocks:
  • A lightweight centralized algorithm for resource allocation and flow control, dynamically optimizing performance and costs as a function of traffic changes.
  • A distributed and instantaneous prioritization layer that diverts rejections to unimportant traffic, overcoming overcharges at optimization-interim periods.
• AI based micro-services integrate into the framework.

Value Proposition

• The unique combination of the technology allows achieving cost-optimal operations via the optimization layer, while reacting instantaneously to fluctuations and problems avoiding over charges.
• Technology is backed by patent (pending) and internal know-how.

The Team

• Anat Bremler-Barr. Professor of Computer Science and AI at Tel-Aviv Univ. She was a co-founder and chief scientist of Riverhead Networks, a company providing mitigation  systems from Denial-of-Service attacks, that was acquired by Cisco.  She has twenty years of experience in researching network security and efficiency of networks with dozens of patents.  Prof. Bremler is Founder and director of the Deepness Lab, which focuses on designing reliable and efficient networks and network devices.
• Hanoch Levy. Professor of Computer Science and AI at Tel-Aviv Univ. Tens of years of experience in researching performance of computer and communications networks.  This includes vast experience in industry including large companies as well as startups. Served (twice) as a head of school/department of Computer Science at TAU.
• Jhonatan Tavori. PhD from the School of Computer Science and AI at Tel Aviv Univ, specializing in Networks Security and Optimization. Recipient of the Blavatnik Prize for Outstanding PhDs in Computer Science, the Fulbright Fellowship, Zuckerman Leadership Scholar, and the EU Next Gen Internet Fellowship. Brings years of practical hands-on experience in systems and networks.