Balex Technologies LLC

Accelerating the Speed of Data Flow in  Computers Using the Scalable Data Access System (SDAS)

Technology Summary:
The wait time to move data into and out of computer memory and processors is called latency. Latency is the single biggest performance inhibitor for  I/O (Input/Output) intensive applications such as artificial intelligence, cybersecurity and numerical modeling, causing processors in even the fastest high performance computer systems (HPC s) to spend a significant, if not a majority of their time, idling, waiting for data. Latency not only causes significant delays in the important decisions that rely on these applications, but it is also causes significant waste of resources, including computer run time, the energy to run and cool the computer, as well as staff time.

Although I/O intensive applications that most rely upon fast processors  and ultra-low latency, such as HPC simulations, big data analytics, artificial intelligence, pattern recognition, and cyber security, are widely recognized as having the biggest latency problems, latency also impairs performance in many other systems and applications, such as enterprise and cloud computing, and the development of the internet of things (IoT). Moreover, it is common for programmers and users to attempt to avoid I/O issues by dumbing down their code (making it less rigorous) by, for example, estimating variables rather than calculating them. The dumbing down of software is a major hidden source of computer performance impairment.

Balex Technologies LLC is developing SDAS, patented computer technology (we have 6) demonstrated to accelerate I/O processes and reduce latency by over an order of magnitude. Critically, our technology conforms to SCSI and NVMe storage protocol standards, which means that it works with all existing SCSI and NVMe compliant computer hardware and software. Customers will have no need to replace their existing hardware or rewrite applications. Our technology is hardware with embedded proprietary software. Future generations of hardware will be miniaturized and included in mobile and portable systems. Balex is collaborating with the Texas Advanced Computing Center (TACC), a leader in the deployment and scaling of new technologies.

Market: 
We intend license our SDAS to the suppliers of new multicore computers such as IBM, HP and Dell. In addition, adding SDAS to already existing computers will make them more efficient and faster than when they were new. This will allow businesses who already have them to extend the life of their hardware by several years before they would normally be decommissioned. A full business plan is available for interested investors. 

The Team:
Our inventor, Branislav Radovanovic, has an MS in Computer Engineering from the University of Southern California. He is the inventor of Balex Technology’s Scalable Data Access System (SDAS) and holds 6 patents for this technology. He has spent over 20 years working at major computer technology firms including DDN, Intel, Seagate, IBM, NetApp and Cray, and is currently at NOAA. He has extensive experience in high-performance block and file level storage design and high-performance computing system products, including the design and integration of mainframe and data storage solutions (both hardware and software), system performance evaluation and tuning, high-speed networks, software development and problem troubleshooting. He has very deep knowledge and understanding of SCSI, Fibre Channel, iSCSI, iSER, SAS, NVMe, InfiniBand, SRP, OPA, and 10/40/56/100GbE protocols. In his current position for NOAA, he is a member of the Senior Production Analysts team responsible for production workloads on the Weather and Climate Operational Supercomputing System. 

Dr. Martin Perlmutter is a quantitative earth scientist with a PhD from the University of Miami’s Rosenstiel School of Marine and Atmospheric Science. Over Marty’s 35-year career as senior scientist and technical manager at Texaco, Chevron and Argonne National Laboratory, he was responsible for designing, developing, managing and deploying multi-disciplinary R&D programs, especially those related to AI for basin analysis and oil and gas exploration, paleoclimate and automated geophysical interpretation, and 4D simulations and advanced statistics. He holds or has pending several patents in these areas. Almost all software was designed to run on multicore or high performance computers. His specific technical efforts included utilizing supercomputers to run earth and atmospheric models he helped create and run. He has previously been PI of grants for both DOE and NSF. 

Lawrence London  is a lawyer and accountant. He received his JD from Stanford Univ Law and his CPA from the Wharton School of Finance. He was formerly a partner at Ruben & Aronson, LLP and general counsel and CFO of Edifice Solutions, a design-build construction firm which grew over ten-fold in revenues and size during his 3-year tenure. Currently he is a mentor to startups through George Washington University’s Office of Innovation & Entrepreneurship.