PSPAT or Parallel Scheduling and Parallel Transmission is a new and complete packet scheduling architecture. It is aimed at systems that can generate tens of millions of network packets per second, ex. dedicated Cloud environments that run Virtual Machines. Most hardware and packet schedulers in use today fail to be fully efficient under such high load conditions. PSPAT also happens to be independent of the packet scheduling algorithms that can be used with it. The central idea of PSPAT is to… Show more PSPAT or Parallel Scheduling and Parallel Transmission is a new and complete packet scheduling architecture. It is aimed at systems that can generate tens of millions of network packets per second, ex. dedicated Cloud environments that run Virtual Machines. Most hardware and packet schedulers in use today fail to be fully efficient under such high load conditions. PSPAT also happens to be independent of the packet scheduling algorithms that can be used with it. The central idea of PSPAT is to get rid of queuing structures which use locks and thus often often bottleneck the performance of the entire system, even when running on multiple processors and with links that have high packet transfer rates.Link - https://wiki.freebsd.org/SummerOfCode2018Projects/PSPAT Show less

Most modern multicore systems today are Non Uniform Memory Architecture (NUMA), popular for its characteristics like scalability and low latency. NUMA is a shared memory architecture that describes the placement of main memory modules with respect to processors in a multiprocessor system, designed to surpass the scalability limits of the SMP architecture. In the NUMA architecture, each processor has its own local memory module that it can access directly with a distinctive performance… Show more Most modern multicore systems today are Non Uniform Memory Architecture (NUMA), popular for its characteristics like scalability and low latency. NUMA is a shared memory architecture that describes the placement of main memory modules with respect to processors in a multiprocessor system, designed to surpass the scalability limits of the SMP architecture. In the NUMA architecture, each processor has its own local memory module that it can access directly with a distinctive performance advantage. At the same time, it can also access any memory module belonging to another processor using a shared bus or some cache-coherent high-performance links. Like most other processor architectural feature, ignorance of NUMA can result in sub-par application performance. The main goals of adding NUMA support in MariaDB were to reduce lock contention and maximize memory allocation on local node, in order to harness NUMA’s full potential for maximum portability and transparency for the user. Such considerations included processor affinity, memory allocation using implicit operating system policies, and the use of the system APIs for assigning and migrating memory pages using explicit directives.Link - https://github.com/MariaDB/server/compare/10.5...theGodlessLakra:NUMA_Support Show less

● Created Arduino Projects designed to teach the basics of Robotics to school students for India’s leading Education Company.● Tested UK based Pi-Top products and suggested improvements to suit the target customer’s need.● Tested China based XPartner’s Abilix Robots and concluded them as unsuitable for the Indian Education Market.