Ankit Gaur

Testing and verification for gRPC based SAI APIs for Underlay, BFD and NAT related functionalities. Wrote gRPC client based on the proto definitions to test the APIs.

Test and Automation for FRR based Underlay and BFD support on DPU. Testing corresponding SAI (Switch Abstraction Interface) APIs over gRPC. Automated test cases in pytest based framework. Automated various scenarios including Underlay bring up, BFD triggers etc in way that same suite can be run in simulation and physical environment.

The P4 UI Compiler Output Visualization tool helps customer visualize their entire P4 pipeline and the tables within to help in debugging and optimizing them.

Tested and Qualified support for SONiC on Pensando DPU. DPU underlay services are exposed through SAI APIs and it runs as docker container conforming to SONiC's containerized architecture and interacts with swss, syncd etc core SONiC services. Verified platform, underlay, copp, dpu upgrade etc functionalities.

Qualified increase in number of policies supported on the Pensando SDN controller. Running scale tests using simulators and qualifying micro-segmentation with the improved scale. Automated scale topology deployment using Jenkins.

Qualified support of Vmotion functionality on the Aruba Pensando DSS Smart Switch with Private VLAN (PVLAN). Test planning, topology setup, executing tests and automation.

Took ownership of the pyats based Scale Automation suite for Pensando SDN controller. Stabilized and extended the suite for various test scenarios using pyats. Setup Jenkins job to run the suite on demand.

Pensando DPU is P4 based SmartNIC which provides micro segmentation for ESXi based workloads. This is controlled by k8s based SDN controller called PSM. I qualified "Workload Endpoint Metrics Reporting" feature on this, which provided L4 metrics for the micro segmented workloads such as VMs on the Pensando SDN controller. The project involved metrics generation on DPU using traffic tools such as iPerf, setting up workloads on ESXi, and debugging the control path and data path on the DPU.

Responsible for qualifying support of Cohesity Data Platform Virtual Edition (VE) deployment on various virtualization flavours viz. ESXi, KVM and Hyper-V. Maintained and wrote automation in Go programming language using ginkgo framework.

Developed a generic Python based framework to automatically discover the software capacity by adjusting the load on the machine (on basis of monitored metrics) given the limiting values for various parameters like CPU, RSS, request latency etc.

The MSL4 Live-to-VoD feature enables customers to convert live streams to video-on-demand assets for HLS and DASH. I worked as SDET engineer for the core features which involved indexing the HLS and DASH playlists and saving it on interval key-value store. I wrote parsing utilities for these protocols (by utilizing open source libraries) as well as implemented a key-value store simulator using Python PickleDB library. Technologies: HLS, DASH, Python, Demuxing utilities.

Previously the unittest based test automation framework didn't support re-running just the failed test cases, which is helpful in saving time while qualifying a release. I added support for this by using testid plugin provided by nose framework and integrating it into the automation framework.

Developed a Pre-checkin test system to qualify the code before being committed to the trunk branch. This helped in preventing bad checkins to the master and helped maintain master stability . Used Perforce shelve functionality, Jenkins pipeline, Jenkins Blue Ocean and Consul as key-value store. This also served as introduction to Jenkins 2.0 for the team, and I helped out in Jenkins 2.0 pipeline adoption.

Added crash reporting functionality to the Jenkins based Continous Integration system which gives information about crashes happening in CI along with relevant information to help in debug and fix the issue. Developed this using linux kernel's "core_pattern" functionality in combination with a data collector script.

Tested and qualified support for end to end SSL on the USS VOD Translator deployed in Akamai's network. Tested integration with Akamai KDC (Key Distribution Center) and associated security check. Qualified the resultant performance impact and capacity.

Code coverage for C++ based software with lcov instrumentation and gcov frontend.

Took SDET ownership of USS VOD Translator product, a C++ based server, which supported content transmuxing (from and to Akamai HD Flash, Adobe HDS, and Apple HLS) format for Akamai's HTTP based Video On demand workflow. Supported product features like SSL support, captioning support for WebVTT and did capacity testing to provide performance numbers to ops team for deployment sizing. Also qualified multiple software and Platform pack releases for the component. Moved various cron jobs to Jenkins. Took care of day-to-day responsibilities like debugging automation nightly failures, adding automation using perl based framework.

Co-owned the Jenkins and Ansible based CI setup for the project, which involved adding new projects, enhancing the ansible playbooks etc. Developed understanding and skills in this area. Also helped in reducing the number of machines required for testing by merging two machine roles into one. Helped other teams in the org to ramp-up on advanced aspects of CI. Moved Nightly runs, Install tests and Load/Benchmark tests to Jenkins from Cron jobs - this helped in better managing the runs, provided much more visibility into the runs progress and made it easier for others to run it.

Worked as SDET on the various modules for the Akamai's new solution for Low Latency Streaming called MSL4. Handled various features involving archiving of data, support for various media protocols (like DASH, HDS, HLS etc), and system features.

This was the first feature I worked on, in my role as SDET in Akamai's Media Division. I was involved in test planning and automation (in a Perl based framework) to enable secure end-to-end video delivery over HTTPS for Mid-Tier components. This involved interaction with KMI (Key Management Infrastructure) to obtain the certs and other secrets. I developed fine grained understanding of how automated key management infrastructure is run at the global scale. Technology: HTTPS, SSL, KMI, KDC

The Defensics Codenomicon tool provides automated solution to test the protocol implementation for robustness against malformed packets (like one which would trigger buffer overflow, format string, memory allocation issues). I took initiative to start protocol robustness testing on IronPort Web Security Appliance. Trained and engaged team on effort, Developed strategy and best practices for the tool usage. This helped us identify many of memory leaks and crashes.

Tested and qualified NAT66, NAT46, NAT64 on ASA firewall.

Qualified IPv6 support for Network Address Translation on Cisco Adaptive Security Appliance. Also worked on adding TCL based automation and qualifying the stabilization efforts for the functionality.

SaaS Access control centralizes the management of access to multiple SaaS applications (like SalesForce, Webex, Google Apps) by integrating the the SaaS application authentication with the already existing authentication mechanism in organization (like NTML/LDAP etc). The Web Security appliance uses the Security Assertion Markup Language (SAML) to authorize access to SaaS applications. My role was as test engineer to create test plan, execute the test cases and qualify the feature.

Maintained Python + Selenium IDE based build qualification suite to smoke test the build before taking up further testing. This helped us to discover any basic defects at the start itself, so as to not waste further efforts.

Test planning and execution for feature which added support for configuring Authentication type on per Security Identity basis.

WSA supported NTLM and LDAP authentication to establish identity of the users going via the proxy. I acted as co-owner of the Authentication module from the QA side and qualified various features such as Per Identity Auth, Transparent HTTPS Authentication etc.

This project involved automating the upgrade testing procedure on WSA (Web Security Appliance). The project consisted of modules such as process monitoring, health check, request generation, log parsing etc. For this we used a python based internal framework as well as Expect and Bash scripting. The project halved the upgrade testing time and hence increased efficiency of team.

IronPort Data Security Filters provide administrator with visibility and control over data leaving the network via the web (HTTP/HTTPS) and FTP. This feature allows admin to create policies and take actions based on relevant parameters like the source (user), destination (URL categories and web reputation), and file metadata (file name, file type, and file size). I was involved in doing qa qualification for the feature.

This is a Web Proxy supporting HTTP/HTTPS and later FTP. Joined the IronPort WSA team as QA engineer.

Joined Cisco as NCH (New College Hire) and got assigned to STBU (Security Technology Business Unit) as QA engineer. Attended trainings on CCNA Networking, Bash/Shell Scripting, Testing Methodologies.

Designed and simulation for Digital Filter for NTSC using Matlab. This project involves analysis and prototyping of a noise reduction system decoder in Matlab. It involved study of Analog Televison Processing and its digital implementation in MATLAB.

Brought up Asterisk PBX Based VOIP system to support inbound VOIP calls to campus PSTN telephony network using Digium card.