Rahul Jadhav

Rahul Jadhav

Bengaluru, Karnataka, India
3K followers 500+ connections

Join to view profile

About

Decades of experience in networks, security, transport, engineering and software…

Activity

Join now to see all activity

Experience

  • Accuknox Graphic

    Accuknox

    Bengaluru, Karnataka, India

  • -

  • -

  • -

  • -

    Bengaluru Area, India

  • -

    Mumbai Area, India

  • -

    Mumbai

  • -

Education

  • University of Mumbai Graphic

    University of Mumbai

    -

    Activities and Societies: Thesis work done in IIT Powai under the supervision of Dr. S.S.S.P. Rao in SPARC Labs.

Volunteer Experience

  • Founder, Lead Author

    Whitefield Open Source IoT Simulation Framework

    - Present 8 years 7 months

    Science and Technology

    Hooked on to this idea during my involvement in IETF standardization work and partly because of the deployment issues we were facing with respect to AMI networks. The aim is to solve a developer dilemma, "Will the network stack scale to the size it is anticipated with realistic wireless conditions? What range can be expected? What packet delivery rate can be expected?".

    https://github.com/whitefield-framework/whitefield

  • Open Source Graphic

    Contributor

    Open Source

    Science and Technology

    Contributed to several open sources including, Cilium, Hubble, Scapy, ls-quic, RIOT-OS, Contiki-OS, Linux kernel.
    For details: http://nyrahul.github.io/about/

  • Evangelist, Maintainer

    KubeArmor - Open Source

    - Present 4 years 2 months

    Science and Technology

    KubeArmor is a system policy auditing and enforcement engine for containerized workloads. It gels well with kubernetes environment and provides rich controls.
    Internally, it uses Linux LSMs (Linux Security Modules) such as AppArmor, SELinux to enforce policies. The audit information is correlated with k8s metadata using a mix k8s operator and eBPF byte-code.

    For more: https://github.com/kubearmor/KubeArmor/

Publications

  • Revisiting Acknowledgment Mechanism for Transport Control: Modeling, Analysis, and Implementation

    IEEE/ACM Transactions on Networking

    The shared nature of the wireless medium induces contention between data transport and backward signaling, such as acknowledgment. The current way of TCP acknowledgment induces control overhead which is counter-productive for TCP performance especially in wireless local area network (WLAN)
    scenarios. In this paper, we present a new acknowledgment called TACK (“Tame ACK”), as well as its TCP implementation TCP-TACK. TACK seeks to minimize ACK frequency, which is exactly what is required by…

    The shared nature of the wireless medium induces contention between data transport and backward signaling, such as acknowledgment. The current way of TCP acknowledgment induces control overhead which is counter-productive for TCP performance especially in wireless local area network (WLAN)
    scenarios. In this paper, we present a new acknowledgment called TACK (“Tame ACK”), as well as its TCP implementation TCP-TACK. TACK seeks to minimize ACK frequency, which is exactly what is required by transport. TCP-TACK works on top of commodity WLAN, delivering high wireless transport goodput with minimal control overhead in the form of ACKs, without any hardware modification. Evaluation results show that TCP-TACK achieves significant advantages over legacy TCP in WLAN scenarios due to less contention between data packets and ACKs. Specifically, TCP-TACK reduces over 90% of ACKs and also obtains an improvement of up to 28% on goodput. A TACK-based protocol is a good replacement of the legacy TCP to compensate for scenarios where the acknowledgment overhead is non-negligible.

    See publication

  • TACK: Improving Wireless Transport Performance by Taming Acknowledgments

    SIGCOMM 2020

    We present a new acknowledgment called TACK (“Tame ACK”), as well as its TCP implementation
    TCP-TACK. TCP-TACK works on top of commodity WLAN, delivering high wireless transport goodput with minimal control overhead in the form of ACKs, without any hardware modification. To minimize ACK frequency, TACK abandons the legacy received-packet-driven ACK. Instead, it balances byte-counting ACK and periodic ACK so as to achieve a controlled ACK frequency. Evaluation results show that TCP-TACK…

    We present a new acknowledgment called TACK (“Tame ACK”), as well as its TCP implementation
    TCP-TACK. TCP-TACK works on top of commodity WLAN, delivering high wireless transport goodput with minimal control overhead in the form of ACKs, without any hardware modification. To minimize ACK frequency, TACK abandons the legacy received-packet-driven ACK. Instead, it balances byte-counting ACK and periodic ACK so as to achieve a controlled ACK frequency. Evaluation results show that TCP-TACK achieves significant advantages over legacy TCP in WLAN scenarios due to less contention between data packets and ACKs. Specifically, TCP-TACK reduces over 90% of
    ACKs and also obtains an improvement of ∼ 28% on goodput. We further find it performs equally well as high-speed TCP variants in wide area network (WAN) scenarios, this is attributed to the advancements of the TACK-based protocol design in loss recovery, round-trip timing, and send rate
    control.

    See publication

  • 6lowpan Fragment Forwarding and Recovery

    IEEE Communications Standards Magazine

    We focus on the fragment forwarding techniques of 6LoWPAN when considering IPv6 fragmented datagrams in multihop networks. After providing an overview of the main operations that are specified in RFC 4944 such as fragmentation, reassembly, and fragment forwarding, we list the potential problems that the fragment forwarding mechanism may introduce in multihop networks. Our study shows that RFC
    4944 comes with a number of issues in terms of network reliability, end-to-end delay, resource…

    We focus on the fragment forwarding techniques of 6LoWPAN when considering IPv6 fragmented datagrams in multihop networks. After providing an overview of the main operations that are specified in RFC 4944 such as fragmentation, reassembly, and fragment forwarding, we list the potential problems that the fragment forwarding mechanism may introduce in multihop networks. Our study shows that RFC
    4944 comes with a number of issues in terms of network reliability, end-to-end delay, resource usage, and implementation. Then we introduce the latest proposed solutions from the standardization community (i.e., 6lo and lwig WGs) and discuss their drawbacks.

    See publication

  • A new face descriptor using local un-quantized patterns

    IEEE - Signal Processing Image Processing & Pattern Recognition (ICSIPR), 2013 International Conference

    We present a novel face representation based on local un-quantized patterns (LUP) descriptors. LUP descriptor is a simple yet powerful descriptor which measures the difference of intensities between surrounding pixel with the center in a local neighborhood, but preserves the finer local geometric structure unlike LBP, SIFT or HOG (which uses either the quantized version of local gray level patterns or quantized codes of image gradients). This descriptor also solves the problem of limited…

    We present a novel face representation based on local un-quantized patterns (LUP) descriptors. LUP descriptor is a simple yet powerful descriptor which measures the difference of intensities between surrounding pixel with the center in a local neighborhood, but preserves the finer local geometric structure unlike LBP, SIFT or HOG (which uses either the quantized version of local gray level patterns or quantized codes of image gradients). This descriptor also solves the problem of limited spatial support of LBP like operators, where increasing the size of local-neighborhood increases the histogram dimensions exponentially making it unsuitable for real-time needs. By applying principal component analysis (PCA) to LUP, we develop a new srepresentation, which gives better performance than LBP and comparable performance to LARK while only taking a fraction of the computation when compared to the latter.

    Other authors
    See publication

  • Efficient Route Invalidation (RFC 9009)

    IETF

    The document explains the problems associated with the use of No-Path Destination Advertisement Object (NPDAO) messaging in RFC 6550 and also discusses the requirements for an optimized route invalidation messaging scheme. Further, this document specifies a new proactive route invalidation message called the "Destination Cleanup Object" (DCO), which fulfills requirements for optimized route invalidation messaging.

    See publication

  • Improved syscall batching for network I/O

    Netdev 0x13 Conf

    Network applications can benefit from reduced CPU cycles by amortizing the system call overhead of
    network I/O operations.
    We review two existing interfaces for network I/O batching namely recvmmsg()/sendmmsg() and SO_RCVLOWAT and propose extensions to these mechanisms.
    We found an 8x syscall reduction with our traffic patterns (gaming scenarios) with the usage of such
    extensions.

    See publication

Patents

  • METHOD FOR PERFORMING IN-KERNEL INSTRUMENTATION OF KERNEL-LEVEL EVENTS

    Issued US 2024/0086558 A1

  • Network switching method, electronic device, and system on chip

    Issued US11910489B2

    A network switching method includes transmitting, by an electronic device, data using a wireless local area network, when a data transmission delay of the wireless local area network is greater than a first threshold, switching, by the electronic device, to transmitting data using both the wireless local area network and a mobile network, and when the electronic device transmits data using both the wireless local area network and the mobile network, if the data transmission delay of the…

    A network switching method includes transmitting, by an electronic device, data using a wireless local area network, when a data transmission delay of the wireless local area network is greater than a first threshold, switching, by the electronic device, to transmitting data using both the wireless local area network and a mobile network, and when the electronic device transmits data using both the wireless local area network and the mobile network, if the data transmission delay of the wireless local area network is greater than a second threshold, switching, by the electronic device, to transmitting data using the mobile network, where the first threshold is less than the second threshold.

    See patent

  • Data Processing Method and Apparatus

    Issued US 17/097,565

    In my words:
    The idea optimizes batch processing of packets using in-kernel optimization techniques considering multi-path transports.

    Abstract:
    A data processing method and apparatus, where one or more applications run in an operating system of a first device, there is a plurality of paths between the first device and a second device, and the paths correspond to the one or more applications. In a data receiving process, the one or more applications obtain data on the paths through…

    In my words:
    The idea optimizes batch processing of packets using in-kernel optimization techniques considering multi-path transports.

    Abstract:
    A data processing method and apparatus, where one or more applications run in an operating system of a first device, there is a plurality of paths between the first device and a second device, and the paths correspond to the one or more applications. In a data receiving process, the one or more applications obtain data on the paths through a system call, and further obtain data in a plurality of packets from at least one path. In a data sending process, the one or more applications distribute data to corresponding paths through a system call.

    See patent

  • DEVICE AND METHOD FOR BALANCED AD-HOC NETWORK FORMATION

    Issued OO PCT/CN2017/070167

    In simple words:
    The idea is about balancing the RPL network formation using various metrics and how these metrics are signaled in the overall network.

    In one implementation, the present invention provides a mechanism for balanced ad-hoc network formation. To achieve the for balanced ad-hoc network formation, the present invention sends the metric information with DIO control message. A new metric container type is introduced in RPL to hold the metric information, and select the…

    In simple words:
    The idea is about balancing the RPL network formation using various metrics and how these metrics are signaled in the overall network.

    In one implementation, the present invention provides a mechanism for balanced ad-hoc network formation. To achieve the for balanced ad-hoc network formation, the present invention sends the metric information with DIO control message. A new metric container type is introduced in RPL to hold the metric information, and select the parent with minimum path cost, and switch the parent with minimum path cost, as per the defined logic.A stateless metric considers that the average traffic flow from each node is approximately same. The stateless metric directly uses the number of routing table entries (RTsize) to decide which path to use. A stateful metric, wherein the node keeps a track of packets processed per second by itself. For stateful metric, the node maintains packets processed rate (PPR) variable, which is then used as the metric.

    See patent

  • Optimized data reception using opportunistic time bound batch processing

    Filed US 86211071US04

  • Method for nodes identification and allocation in Wireless Mesh Networks

    Filed CN 86730578CN01

    Way to uniquely allocate 2-byte node id in multihop mesh networks and further use it for 1) at the link layer for optimizing channel selection 2) randomizing data send time.
    The node ids could be federated across multiple border routers shared by a common backhaul network.

  • Multi-socket system interfaces to reduce syscall overhead

    Filed CN 86211082CN01

    Aggregating send/recv calls across multiple sockets for reducing system calls overhead without incurring use of mmapped-IO.

  • Network Switching Method, Electronic Device, and System On Chip

    Filed US US2020245406A1

    Embodiments of this application provide a network switching method, a system on chip, and an electronic device, to improve user experience and reduce unnecessary switching overheads in a scenario of transmitting data based on a plurality of types of wireless networks.

  • Method to lower data transmitting delay using multiple path

    Filed CN 85911166CN01

    Mobile application types are becoming more and more abundant, and some new mobile applications have higher demands for business interaction delay performance. The wifi network and the LTE network supported by the terminal can enhance the service interaction delay performance through complementary packages. Need a mechanism to guarantee user experience under the premise of low network overhead.
    The user's QoS indicator is used as the network packet state control threshold, and the…

    Mobile application types are becoming more and more abundant, and some new mobile applications have higher demands for business interaction delay performance. The wifi network and the LTE network supported by the terminal can enhance the service interaction delay performance through complementary packages. Need a mechanism to guarantee user experience under the premise of low network overhead.
    The user's QoS indicator is used as the network packet state control threshold, and the corresponding network packet state transition state machine is designed to ensure the user experience under the premise of low overhead cost;

  • Real time video stream optimization for TCP/HTTP transport

    Filed CN PCT/CN2019/073922

    Optimization of real-time video streams on TCP by managing buffers appropriately in the network and the kernels. Prioritizing in-buffer control and video stream frames.

  • Method of Conditional Access Service Provisioning for Service Level Negotiations for Constrained Devices

    Filed OO PCT/CN2016/101498

  • Bandwidth management device, central management device and method of bandwidth management

    Issued US US 20150106480 A1

    Other inventors
    • liubingzhen
    See patent

  • Method for estimating face bounding box for high performance face detection

    Issued EU WO/2016/019709

    Other inventors
    See patent

  • COMPUTER-IMPLEMENTED SYSTEM AND METHOD FOR PERFORMING IN-KERNEL INSTRUMENTATION OF KERNEL-LEVEL EVENTS

    Filed 17/941,938

    In kernel aggregation of visibility/observability events to reduce performance impact of context switching. The aim is to do aggregation at the earliest (in-kernel) in the event life cycle.

  • SYSTEM AND METHOD FOR SPECIFYING AND MANAGING PREDEFINED POLICIES FOR CONTAINERIZED WORKLOADS

    Filed 17/689,981

  • Proximity based distributed caching of Service information in constrained multi-hop networks

    Filed EU WO2017041631A1

    Communication networks devices and methods for proximity based distributed caching of service information within said network are disclosed. The network comprises service endpoint device (1402), client devices (1404), resource directory devices (1406), and border router devices (1408), in communication with each other. The methods for proximity based distributed caching of service information comprises: identifying (1502) RDs in a network; identifying (1504) service endpoints or clients in the…

    Communication networks devices and methods for proximity based distributed caching of service information within said network are disclosed. The network comprises service endpoint device (1402), client devices (1404), resource directory devices (1406), and border router devices (1408), in communication with each other. The methods for proximity based distributed caching of service information comprises: identifying (1502) RDs in a network; identifying (1504) service endpoints or clients in the network and assigning (1506) preferred RD to every service endpoint or client; registration (1508) of service endpoints with the service information to the preferred RD, the RD then backups the information on the BR; client always does the service lookup (service lookup from the client) (1510) from its preferred RD, if that RD cannot locally fulfill the lookup then it queries the BR for information, and based on the lookup patterns the RD decides to cache the service information i.e., proximity-based caching (1512).

  • SYSTEM, METHOD AND NODE FOR TIME SYNCHRONIZATION USING RPL PROTOCOL

    IN IN201641044489

    The present disclosure discloses system, method and node for time synchronization using RPL protocol. In contrast to the prior-art techniques, the present disclosure by system, method and node for time synchronization without incurring additional messaging overhead and achieve synchronization efficiency at par with existing techniques by reusing LLNs routing protocol primitives. The RPL is utilized to handle time synchronization as part of network formation which is referred to as Time…

    The present disclosure discloses system, method and node for time synchronization using RPL protocol. In contrast to the prior-art techniques, the present disclosure by system, method and node for time synchronization without incurring additional messaging overhead and achieve synchronization efficiency at par with existing techniques by reusing LLNs routing protocol primitives. The RPL is utilized to handle time synchronization as part of network formation which is referred to as Time Synchronization protocol using RPL for adhoc wireless sensor network (TSRPL) in this document. In order to sync time with every wireless sensor node, a non-loop hierarchical structure is required to propagate time synchronization across network, this requirement is satisfied as the RPL serves purpose to build a multi-hop tree like (DAG) network among sensor nodes, and timing handshake is performed using DAO- DAO/ACK by piggy backing time handshake information of sender-receiver approach of timing synchronization.

  • A RELAY NODE AND METHOD FOR TRANSPARENT RELAYING OF PACKETS USING STATEFUL MEDIA ACCESS CONTROL (MAC) LAYER OF WIRELESS SENSOR NETWORK

    IN IN201741011060

  • Optimized route invalidation using modified No-Path DAO (NPDAO) signaling

    IN WO2018113394A1, IN201641043305

    Apparatus and method for optimized route invalidation using modified no-path DAO signaling are disclosed. According to the present invention, a node switching its current parent is adapted to send a regular Destination Oriented Directed Acyclic Graph (DODAG) Advertisement Object (DAO) message. Using the changed signaling, the common ancestor node generates an NPDAO on behalf of the switching node on receiving a refreshed DAO from an alternate path. The common ancestor node according to the…

    Apparatus and method for optimized route invalidation using modified no-path DAO signaling are disclosed. According to the present invention, a node switching its current parent is adapted to send a regular Destination Oriented Directed Acyclic Graph (DODAG) Advertisement Object (DAO) message. Using the changed signaling, the common ancestor node generates an NPDAO on behalf of the switching node on receiving a refreshed DAO from an alternate path. The common ancestor node according to the present invention reuses a same PathSequence from the regular DAO based on which the NPDAO gets generated. The common ancestor node detects routing anomaly using next hop mismatch on reception of the DAO to generate the NPDAO on behalf of the target node. The No-Path DAO traverses downward/downstream along the previous path, unlike any other DAO messages traverses upstream according to convention approaches. The present invention provides a new signaling mechanism for route invalidation in constrained networks.

  • REAL TIME OCCLUDED FACE DETECTION IN MULTIMEDIA OBJECTS

    IN IN201641013481

    The present application refers a method for real time face detection in multimedia objects and apparatus. In line with the present invention,the type of occlusion in the occluded multimedia object is classified using a classifcation mechanism. Thereafter,a non-occluded multimedia object is reconstructed using an image reconstruction technique based on the identified type of occlusion and the reconstructed multimedia object is validated.

Projects

  • KubeArmor

    Cloud-native Runtime Security Enforcement System

    See project

  • Optimized Mesh solution for low power Wifi for Home/Building Automation

    - Present

    Deployment of wifi based home automation solutions is much easier than its counter-parts such as LoWPAN, PLC, BLE. However, the challenge is power efficiency and performance. We envisaged a design for using low-power wifi (with high RDC nodes and power management features from 802.11s) and in many ways it is different from the competitors. The recent advancements in Wifi physical/L2 layer (especially the Wifi-NAN) seems conducive to such networks.
    As a lead designer and core module developer…

    Deployment of wifi based home automation solutions is much easier than its counter-parts such as LoWPAN, PLC, BLE. However, the challenge is power efficiency and performance. We envisaged a design for using low-power wifi (with high RDC nodes and power management features from 802.11s) and in many ways it is different from the competitors. The recent advancements in Wifi physical/L2 layer (especially the Wifi-NAN) seems conducive to such networks.
    As a lead designer and core module developer it is a challenge to work with unknowns and design a multi-hop mesh network solution at L3 with legacy IPv4/IPv6 non-mesh station.

  • Transport Optimization for terminals

    - Present

    Transport optimization for upstream video streaming based on frame prioritization and managing TCP buffer occupancy.
    Throughput optimization for shorter HTTP flows by improving throughput in TCP slow start phase.
    Worked on various congestion control algorithms such as CUBIC, RENO, BBR and proposed various techniques to improve upon these algorithms with respect to the given conditions.
    Worked on prototyping on Android/Linux and involved changing Linux kernel code and Android framework…

    Transport optimization for upstream video streaming based on frame prioritization and managing TCP buffer occupancy.
    Throughput optimization for shorter HTTP flows by improving throughput in TCP slow start phase.
    Worked on various congestion control algorithms such as CUBIC, RENO, BBR and proposed various techniques to improve upon these algorithms with respect to the given conditions.
    Worked on prototyping on Android/Linux and involved changing Linux kernel code and Android framework code.

  • Whitefield-Framework (An Open Source Framework)

    - Present

    Whitefield aims to provide a simulation environment for wireless sensor networks by combining RF simulation provided by NS3, Castalia/Omnet++ while allowing to use native mode configuration of popular IoT OSes such as Contiki and RiOT.

    See project

  • Optimized mesh network for Smart Meter based on 802.15.4

    -

    Optimized protocol stack solution for Smart Meter IoT networks. Worked on optimizing routing protocol to improve convergence time and repair time. Worked on scaling the network to more than 1500 nodes with constrained RF such as 802.15.4. Security solutions for constrained network based EAP-PANA-PSK/TLS with compressed signalling to adapt to lower MTU LoWPAN networks.

    Standardization activities in this…

    Optimized protocol stack solution for Smart Meter IoT networks. Worked on optimizing routing protocol to improve convergence time and repair time. Worked on scaling the network to more than 1500 nodes with constrained RF such as 802.15.4. Security solutions for constrained network based EAP-PANA-PSK/TLS with compressed signalling to adapt to lower MTU LoWPAN networks.

    Standardization activities in this context:
    https://tools.ietf.org/html/draft-ietf-roll-efficient-npdao-00
    https://datatracker.ietf.org/doc/draft-jadhav-lwig-nbr-mgmt-policy/

    Worked towards open source solutions such as RIOT, Contiki, LWIP.

  • Video Streaming Optimization for terminals/enterprises

    -

    -Worked on ideas to reduce video stream size (for Youtube, Youku) by light weight H264 stream transformation.
    - Bit rate reduction by dropping unreferenced frames. Further frame prioritization scheme which worked by calculating weight of each frame by taking into account macro-block referencing was implemented.
    - CAVLC to CABAC lossless conversion
    - Implemented Casacaded Pixel Domain transcoder (CPDT) for higher bit-rate gains but with relatively higher computational…

    -Worked on ideas to reduce video stream size (for Youtube, Youku) by light weight H264 stream transformation.
    - Bit rate reduction by dropping unreferenced frames. Further frame prioritization scheme which worked by calculating weight of each frame by taking into account macro-block referencing was implemented.
    - CAVLC to CABAC lossless conversion
    - Implemented Casacaded Pixel Domain transcoder (CPDT) for higher bit-rate gains but with relatively higher computational usage.

    Coupled video domain expertise with transport optimization to achieve improved QoE for VoD on Huawei handsets.

Languages

  • Chinese

    Limited working proficiency

  • Marathi

    Native or bilingual proficiency

  • Hindi

    Full professional proficiency

  • English

    Full professional proficiency

Recommendations received

1 person has recommended Rahul

Join now to view

More activity by Rahul

View Rahul’s full profile

  • See who you know in common
  • Get introduced
  • Contact Rahul directly
Join to view full profile

Other similar profiles

Explore more posts

Explore collaborative articles

We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.

Explore More

Others named Rahul Jadhav in India

3228 others named Rahul Jadhav in India are on LinkedIn

See others named Rahul Jadhav

Add new skills with these courses

See all courses