SRIDHAR IYER

ENGINEERING LEADER / PHOTOGRAPHER / RUNNER

I consider myself a versatile technologist with industry experience in various networking technologies.. and some experience in shooting events as a photographer.

  • Sridhar V Iyer

Employment

  • Senior ArchitectJuly'14 - Present

    Versa Networks
    • Leading the adoption of ML/AI & cloud deployments within different products.
    • Designed steering of SAAS traffic based on application performance metrics.
    • Designed & implemented the Acctmgr module for Versa FlexVNF, which maintains lock-free, high-frequency, high performance counters across multiple processes & threads, backed up by REDIS db.
    • Designed & implemented cos-shaping and Adaptive CoS shaping for Versa FlexVNF solutions. This was developed using intel DPDK to perform the actual rate-limiting, and BGP to convey the message to different clients.
    • Implemented 802.3ah EOAM/CFM (Connectivity fault management subsystem).
    • Implemented AE(Aggregated Ethernet), static and lacp(802.1ax) based, for Versa FlexVNF. Proprietary extensions were added to support non-standard topologies.
    • Designed & Implemented various infrastructural elements(SKU limit handling, licensing etc) across various teams and products.
  • Sr Software EngineerJuly'10 - July'14

    Cisco Systems Inc.
    • Involved in the development of the first Cisco n9000 series of switches.
    • Involved in the development of the new Cisco n6000 series of switches. Worked on the forwarding layer right from the DOL stage to the release stage. This entailed validating the new asic model (based on RTL code), modifying the existing OS to work with the updated hardware and registers, implementing new forwarding related features and porting existing features to the new hardware. A lot of work was also done troubleshooting the data flows on the switch. This switch would be the first Cisco NXOS switch to be used in the spine layer, between the access and aggregation layer.
    • Develop and maintain the l2 layer of the forwarding plane on Cisco's n5000 series of switches.
    • Developed and maintained various authentication/authorization/monitoring modules on n5000 switches like SNMP, AAA, TACACS, RADIUS etc.
  • Software EngineerJune'07 - July'10

    Mailshell Inc.
    • Designed, wrote and maintained tools used for anti-spam analysis. These ranged from simple Bayesian classifiers to multi-core proprietary engines & enhanced the existing proprietary Anti-spam SDK and developed other tools required by the clients (such as proxies and plugins to use the SDK).
    • Spearheaded the adoption of companywide python usage for regression tests and continuous integration. This lead to a reduction of test development time by a factor of ten and almost doubled the number of regression tests by a factor of three.
  • Research AssistantAugust'06 - May'07

    Dept Of Computer Science, Syracuse University
    Under Dr Wenliang Du
    • Developed instructional laboratory projects for computer security education. These projects are used by the faculties and students (undergraduate and graduate) at various universities as laboratory exercises for computer system security education. These included demonstrating various attacks like buffer overflow, LD_PRELOAD, return-to-libc, format string vulnerability etc.
    • Developed various security related products on Minix 3.1.2a such as Encrypted File System, IPSec (ESP tunneling with authentication based on RFC 2401 and 2406) and ASLR.
    • Project funded by National Science Foundation ($451,682, 01/2007-12/2010. Grant No. 0618680).
  • Research AssistantSeptember'05 - August'06

    Dept Of Psychology, Syracuse University
    Under Dr Marc Howard
    • Did an Independent study on Multidimensional Function Minimization using Genetic Algorithms. The objective was to research the use of genetic algorithm as a viable alternative to Nelder-Mead for fitting cognitive models.
    • Simulated cognitive models (based on Temporal Context Model developed by Dr Marc Howard and Dr Michael Kahana). These simulations are used to prove the correctness of the neural networked model of the short term memory under various hypothetical situations. Standard C++ was used for implementation.

Education

  • M.S. in Computer and Information ScienceAugust'05 - May'07

    Syracuse University, L.C Smith College of Engineering & Computer Science
  • B.E. in Computer Science and EngineeringAugust'01 - June'05

    Visvesvaraya Technological University, S.J.C Institute of Technology

Patents

  • Maintaining Lock-Free, High Frequency, High Performance Counters in Software. US9477471B1
  • Method for configuring a connection using redundant service nodes. [Pending US15091388]
  • Method and computing device for shaping traffic across a wide area network [Pending US15148313]
  • Method and system for shaping traffic from an egress port in a software-defined wide area network [Pending US15814296]
  • Application performance based path-selection using dynamic metrics [Prov US 62686003]
  • Application performance based path-selection [Prov US 62657817]

Contact info

  • Twitter:@codeyman
  • Website: www.sridharv.net