About

199x – 2012

  1. background
    • born in taiwan
    • served in army as constrution engineer/also supports network and backup infrastructure – 1996-98 – retired as sergeant first class
    • studied in USC 1999-2003
    • first job in AT&T lab in 2002 while in school
    • second second with WebMD since September 2004
    • fluent in english and mandarin
    • annual presentation in conferences HP, serena and velocity (web performance) since 2010
  2. current development projects (enterprise software prototyping, continuous integration and technology challenges)
    • continuous integration
      •  8-people task force – manager and director level across QA/PMO/OPS/Dev
      • different roles, PM, architect and technology lead
      • streamline and automate  build delivery/deployment process
      • expect to speed up product delivery pipeline by 60% and reduce communication overhead by 40%
      • currently in purchasing phase – SOW/SLA
    • business dashboard – BI team and business
      • tibco spotfire as analysis/graph engine
      • critical data sources – appfigure, itunes, google, internal personalization database (registration/news letter) and adobe omniture
      • adapters written in java
      • on display in major offices – NYC1/3, ATL and PORTLAND
    • technology intelligence dashboard
    • high impact performance degradation
      • architecture review
      • system/product reliability/performance optimization
  3. management principles
  4. technologies – php/java/unix/database (mysql, oracle/ms sql)

####################################

feedback questions

  • do you think i will be a good candidate for bridgewater associates? if not, how do i improve
  • i read about some negative feedback from bridgewater. can you share with me what your experience is with brigewater?

####################################

mistake i make (technology-related or management-related?)

##################################

process simplification

  • study current process
  • research other unsuccessful cases to avoid making the same/old mistakes
  • design new solution
  • test, test, and test
  • take challenge by smart people

 

2012-14

  1. SDLC efficiency
  2. code analysis – static (model checking and data-flow analysis) and dynamic (how code behaves)
  3. nightly load testing in prod – history replay
  4. no full GC during the week. barely multi-threading, simple coding

metrics analysis and dashboard development

  • employee dashboard – interesting metrics such as page views visits, registrations, user logins
  • tech dashboard –
  • marketing dashboard
  • metrics correlation

incident analysis about critical outages and performance degradation

  • indepth troubleshooting and resolution
  • espeically architecture and platform

Agile and continuous integration task force

  • cross functional team consist of 12 members, directos, architects across dev/ops/sysadmin/qa/pmo
  • new solution implementation
  • simplify, standardize and automate SDLC process

new technology/solution – research and evaluation

About:

  1. senior director
  2. 3 architects
    • reliability/performance engineering
    • CI (dev/ops/qa group of 15) and SDLC efficiency
    • tools and standards
    • metrics: KPI, depth of ‘purepath’; benchmark and metrics standard. SLA. definition of ‘good’
    • ad-hoc troubleshooting.
    • dashboard development
  3. attend conferences about tools/operations/performance – one presentation/year since 2010
  4. voluntary works – vice principal and IT director
    • java to php
    • from self and expesnive hosting to AWS
  5. vintage audio hobbyist

Troubleshooting Procedures

Collect background information

  • Scope of impact
  • Information about the product(s) impacted
  • Interview personnel involved

Initial interpretation

  • Type of incident – outage, service degradation
  • Expectation from senior management
  • Depth and scope of investigation
  • Resource planning

In-depth analysis

  • Timeline analysis
  • Change analysis
  • Log analysis
  • Monitoring data correlation

Research

  • Vendor documentation and white paper
  • Architecture review
  • Code review and application profiling
  • Infrastructure review

Resolution and recommendation

  • process
  • design and architecture
  • infrastructure

use case – cnn slowness

  1. ping test/traceroute of cnn.com
  2. browsers add-ons and plugins
  3. state of PC – i/o, cpu and memory consumption
  4. router issue
  5. cable modem issue

Velocity 2013 Proposal (not accepted)

Proposal: “From hours to minutes – application deployment automation in WebMD

Submitter: Derek Chang (dchang@webmd.net)

Speakers

Rajeev Borborah (rborborah@webmd.net)
Derek Chang (dchang@webmd.net)

Description

Application deployment has been considered a bottleneck in our SDLC pipeline. In this talk we will share our story about achieving deployment automation in WebMD: from case studies, identifying requirements, process simplification and standardization, tool evaluation/selection to implementation. We’ll finish with a live demo – how we use the automation system to promote the code across SDLC.

Session type: 40-minute session
Topics: Operations

Abstract

WebMD started a CI (Continuous Integration) initiative back in October 2011 to deal with the excessive demand for application deployment as a result of:

  • Increasing number of applications supported
  • Diversity of the infrastructure
  • Majority of release pipelines adopt agile approach

This presentation is meant to help organizations like us to prepare and deal with challenges by sharing our complete story with them.

  1. History lesson – how deployment process evolves in WebMD; from ‘cvs -d’ to python scripts, big name deployment tool to what we have today
  2. What we are facing – diversity of infrastructure, number of applications/agile approach which yields high demand for application deployment, configuration management, complexity of deployment procedures
  3. Case studies – deployment procedures as of 2012 and corresponding time analysis and maintenance cost
  4. CI (Continuous Integration) team in WebMD – how we organize a cross-functional task force
  5. Leadership and guidelines How we deal with politics and meet everyone’s need including QA, Dev, Ops, and PMO
  6. Requirement collection – identify features that we need the most: Cost efficiency, configuration management, fail fast, repeatability, Orchestration, Usability, Integration
  7. Vendor selection and evaluation – how do we choose from 20+ potential candidates
  8. Process Standardization and simplification
  9. Culture change and Collaboration – how Dev/Ops/QA team up to implement the system we have today
  10. Before and after comparison – quantify how much we improve
  11. challenges and resolutions – same challenges that YOU might have and how we deal with it
  12. Demo – how we use the automation system to promote the code across SDLC
  13. Q/A

 

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Blog at WordPress.com.

Up ↑

%d bloggers like this: