ravikirandotks@gmail.com | +91.988.682.8569 | https://www.linkedin.com/in/ravikiranks/ | https://github.com/rkks
Ravi is a programmer, author, and consultant residing in Bangalore, IN. Over 20 years of hands-on software development career, he has delivered over #9 products from concept-to-deployment for companies like Versa, Cisco, Juniper and CCPU. In academia, he ranked #1 with 81% aggregate in Masters under the faculty of Computer Science at the RD University, Jabalpur. He pens papers on tech topics, contributes to open-source, and hacks tech in spare time. He is awarded #11 times for timely product delivery, solving critical problems, and help in securing customer design wins. He also provides freelance consultation to startups on need basis.
Industry Trend Analysis . Fire-Fighting . RFP . Customer Engagement . Patent . Articles . Tech Talk . Mentoring . Demos . Product Ownership . Feasibility Analysis . Standards tracking . Cross Team/Geography Collaboration . Distributed System . SD-WAN . SASE . DC Network . Load Balancer . High Availability . Container . NFV . Infra . VOS . NXOS . JUNOS . TWAMP . DPI . VNTAG . VRRP . SCTP . MLAG . LLDP . STP . DPDK . Bonding Driver . C . Bash . Python . Java . Automation . Linux/FreeBSD Network Stack . VFS . Performance . DPU Offload . Ixia . Broadcom Trident3 . Netlogic XLP/R . Fulcrum Bali . Cavium Octeon . MSPD Comcerto 2K . Proprietary ASICs . Docker . LXC . 6Wind FastPath . QNX . Spirent . Solaris . Mirroring File-System . Trillium Stack . SAF .
As a Lead, Ravi delivers #2 large scale solutions, #1 patent on tehnologically significant area, and assists in fire-fighting various customer issues.
Clustering for I/O and Services Scaling over Versa VOS provides an elastic solution to increase aggregate throughput of any deployment by horizontal scaling of VOS nodes. When extended to public Clouds, this can scale on-demand. Aim is to provide a linear scale-out/scale-in architecture to deliver 20 times the throughput of any individual appliance. Ravi implements core algorithm for traffic distribution and session accounting.
TWAMP feature aims to provide fully RFC 5357 compliant, interoperable, and scalable Control/Test protocol stacks for SLA measurement in brownfield deployments having multi-vendor appliances. It interoperates with Cisco IOS, Juniper JUNOS, Nokia AOS and Arista EOS. It scales upto 64K sessions and performance of <20ms inter packet gap. Together with the Versa SASE Client (Endpoint Agent) and Versa Cloud Gateway, it provides last-mile SLA reports with real packet-loss measurements for remote access users and provides historical data. Ravi single-handedly implements whole protocol stack on VOS and ports it to multiple OSes for pairing with respective endpoint agents.
Split-Brain avoidance in Inter-Chassis HA solution provides a novel way to reliably detect peer node down. Solution is granted patent USPTO #10972337. Ravi implements the core solution and guides team members to extend the same. Ravi also implements a solution to detect, isolate, and highlight issues in configuration between nodes participating in Inter-Chassis HA. This assists to resolve issues before they begin to impact functioning of the system. Ravi develops the core framework and provides training, references and help to implement plugins for various HA-aware modules in Versa VOS.
As Individual Contributor, delivers FCoE over Fabric Extender (FEX) integration for proprietary ASIC based F4 line-cards on N7000 family of DC switches and Broadcom Trident3 chipset based N9000 family of DC switches.
Ravi develops innovative lossless on-demand tracing feature for system-wide runtime debug trace logging without affecting performance on NXOS for all N9000, N7000, N3000 family of switches. By nature logs are lossy as buffering, block device driver, and file system syncing work on best-effort basis. Idea is to have elastic, realtime buffer in RAM for lossless log collection. This helps to reduce back-and-forth with customers, improves issue resolution time, and customer satisfaction.
N9000 DC Leaf/Spine Switches form the backbone of ToR & EoR offerings from Cisco for Leaf and Spine Data Center fabrics. Built over ultra-low-cost homegrown ASICs as well as commodity Broadcom chipsets, these switches provide best value for money with consistent low latency and high throughput. As brownfield DCs have a mix of old and new servers, there is demand for higher number of ports for low cost. Fabric Extenders (FEX) fills that gap by extending number of ports available on a switch by many fold at a fraction of their cost. FCoE is a tunneling protocol that provides access to Fiber Channel/FC Storage devices over Ethernet using Converged Network Adapters (CNAs). The ethernet is turned into loss-less SAN through a collection of DC specific standards and extensions like Priority Flow Control, LLDP DCBX, SSPF, Naming Server, and so on. Ravi develops support for FCoE over FEX connected to Broadcom based N9000 line-cards by programming FEX, Trident3 chipset, Control plane, and network drivers on SUP3 as required.
Ravi delivers Linux LXC container based NXOS ISSU support for FCoE N9000 ToR swithces by modifying relevant Kernel Loadable Modules/KLMs in FCoE packet processing path on N9K ToR switches.
N7000 DC switches formed the Core layer of 3-layered DC Fabric architecture of the past. These switches provided high bandwidth, traffic engineering and DC-Interconnect/DCI WAN acceleration features. While some Access layer features like VDP, LLDP/DCBX, User ACLs, and Overlay VTEP were left out in favour of Access switches like N3000, few others like VNTag/FEX, VPC/MLAG, and CDP were implemented in N7000 switches for Collapsed-Core architectures where Core Switch also assumes responsibility of Aggregation layer. Ravi ports support for features like FCoE over Fabric Extender/FEX, FEX Active-Active, vPC/MLAG, Phy-Port vPC, and FabricPath over N7000.
As Individual Contributor, develops control-plane infra for proposed new petabyte scale multi-chassis router stacking project.
As Individual Contributor, develops statistics harvest infra on Netlogic XLP processors, implements clock synchronization mechanisms for SSX3000 Security Exchange Gateway.
As a Technical Leader, delivers #2 carrier-grade and #2 enterprise-grade solutions mentoring a team of 6.
FastPath SCTP is a carrier-grade, highly optimized, stateful gateway solution. It gives bi-di throughput of 1M pps of 1500B SCTP packets over 10G links. It extends 6Wind FastPath stack on NetLogic XLR NPUs. Ravi delivers complex zero-copy packet forwarding & scheduler, in-place IPSec, DMA and DNS support.
Layer2 HA is an enterprise-grade network resilience solution. It gives sub-msec failover, weighted traffic mapping as hub-n-spoke, and faster convergence by VRRP enhancement. It is implemented as extension to Fulcrum ControlPoint on Switch and Bonding Driver enhancement on x86 Blades. Ravi delivers entire product, from ideation to deployment, pens white papers, customer support. It generates >$3M revenue.
FlexBalance is an enterprise-grade server load balancer. It does statistical hashing of traffic marked by L4 flows, MPLS tags, physical port-groups using DWRR & CBQ schedulers. It is implemented as pattern-match & scheduler enhancement using FFU, TCAM. Ravi delivers entire product from code to docs single handedly, works with architect and QA in different geographies, helps in deployment, customer support. Huge commercial success, generated >$8M in revenue, and helped in more than 12 design wins.
LTE ENodeB on Mindspeed ARM processors using 4GMX OS. Mindspeed has ARM family of processors with very low power usage. U-CPU/L-CPU dual CPU achitecture is suitable for distributing Phy and stack layers across two different CPUs, both of which are connected by a message bus. Ravi is tasked with porting, optimization of stack on these low-powered devices
FlexTCA is a carrier-grade, service availability product that pre-integrates Trillium stacks with GoAhead Saffire middleware to run out-of-box on ATCA chassis. It implements OAM interface over SNMP/Web, Compute blade resilience using SAF SAI, Protocol HA using DFTHA layer, Control plane HA using SAF checkpoint service, and Chassis management using SAF HPI. Ravi is responsible for design, code and testing of complex SAF-Trillium integration layer that represents core value-add of product offering.
UpSuite is a high-availability middleware providing fault identification, isolation, and recovery along with real-time mirroring file-system for Solaris based servers. It implements heartbeat framework, application monitoring framework, kernel file-system for mirroring, and NIC resiliency driver. Ravi is involved in doing upgrade of file-system to Solaris 10, supporting zones, and providing bug-fixes on Solaris servers.
As a Technology Consultant, he has helped in feasibility analysis, new tech integration & prototyping/PoC, solving design & coding challenges, timebound resolution of critical bugs, deployment engineering assistance, and Application optimization for scale & performance.
Ravi developed his like for FOSS after he was involved in Linux kernel driver related development activities in 2006. Since then, he has contributed to open-source by identifying & fixing bugs, making own work available as open-source and free under liberal license. Here are some of his early contributions.
Curious coder, who delivers under tight deadlines, automates as hobby, listens to understand, gets job done.
NOTE: This resume is auto-generated using pandoc on markdown text. Latest updated resume is at: html, pdf