Life as Software Engineer Dr. Jeyakesavan Veerasamy University of Texas at Dallas

Agenda Introduce myself Software Engineer: Good & Bad Software: Then & Now SW development environment options SW Efficiency – does it matter? Latest buzz Trends in CS education Open Q&A

Dr. V. Jeyakesavan: Academia, Industry & Personal Dad was a school teacher B.E. (ECE) in CEG Guindy, Anna University – UNIX System Software Engineer, HCL Limited, Chennai, MS Computer Science, University of Texas at Dallas (UTD),

Dr. V. Jeyakesavan: Academia, Industry & Personal … Telecom Software Engineer, Northern Telecom, Dallas, Ph.D. Computer Science (part-time), University of Texas at Dallas (UTD), Technical Lead, Samsung Telecom, Got married in 1998 Adjunct Faculty, UTD CS department, Online Adjunct Faculty in several online universities from 2000

Dr. V. Jeyakesavan: Academia, Industry & Personal … Adjunct Faculty, Southern Methodist University, 2010 Sr. Lecturer (full-time), UTD Computer Science, 2010-present 2 daughters: Nila (8) and Chinmayee (4) Passionate about teaching – happy to share ideas to improve teaching quality in colleges

Dr. V. Jeyakesavan: Summary 18 years experience as Software Engineer 12 years of teaching experience

“Software Engineer” career Great pay! All about logical thinking! Consistently one of the best jobs ! Focus on technical stuff Stay out of politics! Tight deadlines Continuous learning Logical errors Documentation (painful?)

“Software Engineer” career Technical ladder vs. Management ladder Development vs. Test vs. Project management work in big company vs. startup

Complexity of software projects: historical snippets Ad-hoc development model Waterfall model – slow & unpredictable – heavy documentation effort – cost overruns & missed deadlines

My experience with huge SW projects (Waterfall model) Period of instability collisons during merge product delays Is it good enough to release?

Documentation: little vs heavy

flowchart vs. pseudocode

Version control system Enables development by a team of designers

Agile process

Agile methods Limited documentation Better tracking of current status Frequent releases Continuous interaction with the customer Several refinements: XP, SCRUM, …

A typical industry project in 1990 C standard library Home-grown library New code

Same industry project NOW C++/Java standard library Home-grown library Open source components Commercial libraries for industry segment IDE modules New code Follow-up Q: Any disadvantages?

More Questions Expert programmer: what does it mean? Future projects: No code? Just integration? Knowledge of Java/C++/… down the drain?

Application Specific Programming M1M2M3 Intrepreter Mn script … …

Visual Programming Use of visual tools for development, process and testing

Testing 100% testing possible? 100% testing: is it needed? Code coverage tools Automated sanity testing

Software efficiency: Do we really care? Goal is to make software work & meet deadline for most industry projects. Game console: Algorithm takes longer to run  requires higher-end CPU to keep realism  higher price  product fails amid competition Web-server: Algorithm takes longer to run (consider 5 seconds vs. 20 seconds)  tests web-users’ patience & requires more web server capacity. Daily data crunching: What if it takes >1 day?

Data crunching pseudocode initial setup loop – read one tuple – open db connection – send request to db – get response from db – some processing – close db post-processing

Data crunching pseudocode initial setup loop – read one tuple – open db connection – send request to db – get response from db – do some processing – close db post-processing Equation for running time = c 1. n + d 1 Time complexity is O(n)

Data crunching pseudocode initial setup open db connection loop – read one tuple – send request to db – get response from db – do some processing close db post-processing

Data crunching pseudocode initial setup open db connection loop – read one tuple – send request to db – get response from db – some processing close db post-processing Equation for running time = c 2. n + d 2 Time complexity is still O(n), but the constants are different. c 2 < c 1 d 2 > d 1

Concepts: BHCA capacity Busy Hour Call Attempts 1 Million BHCA  central processor should spend < 2.5 milli-second per call (assuming 70% load) Managing BHCA is a “system engineering” activity, done in every software release. Per-call measurements & optimization

Capacity issues 2 types of nodes: – control nodes – transaction processing – CPU load can vary a lot (>60% load is a concern). Use watch- dog timers that automatically reset the node if 100% sustained CPU load is seen. – traffic nodes – actual traffic processing – can safely operate at 90% CPU load Power of Trending

Handling Overload Overload can occur during mega-events or new years day Similar to “Denial of Service” attack Need to shed call requests with minimal effort. Goal is to handle as many requests as possible in reliable manner.

Redundancy Is it for hardware or software? Control nodes: Active/Standby redundancy Traffic nodes: N+1 redundancy Load sharing algorithm? Round-robin or load-balancing Leaky bucket? …

Interesting SW bugs Look at the following code: if (sector_id = 1) … Send call setup message to ALPHA Lab tested the code in alpha sector. What happens when this code is applied to field?

Blocking printf() Debug port used for logs printf() was used to output messages – cannot use break points due to timers CDMA works based on GPS time Timing drift is not good for soft handoffs  handoff failures More time spent in printf()  less time in actual call processing  less capacity

Working with limited pipe There are two types of messages over the air: – Acknowledgement required – No ack required I changed neighbor information message type to improve soft handoff success. Resulted in more handoff failures, since actual handoff processing related messages could not get through.

BSC crashes Unexpectedly long message or spurious content from mobiles causes buffer overrun Fixed size stack was used in OS – more local variables added over time.

Summary: Trends in SW development Code less & integrate more… More use of commercial components/libraries More use of open-source projects Small team & complex projects More acceptance of “Agile process” More automated test cases Focus on Game development & smart-phone applications

Summary: Software career Software will continue to be growth area Requires you to manage your time wisely to maintain quality of life More exciting things yet to come. Put your brain and passion to work! Do not get scared by all the jargons/terms – basics are simple beneath the complex terms! Try to enjoy the job & live a balanced life. Keep emotions away from the job.

University of Texas at Dallas Erik Jonsson School of Engineering and Computer Science Surrounded by 1000s of companies in Dallas- Fort Worth metroplex Students can get internships right after 2 semesters and continue studies in parallel. Most graduate students “graduate debt-free”! We have well-ranked management school too.

Dallas-Ft. Worth metroplex

Indian connection! More than >50% CS faculty are Indian. Around 60% of graduate students are Indian. Management school is named after an Indian: Naveen Jindal School of Management!

40 ________________________________________________________________________ Department of Computer Science Jonsson School of Engineering and Computer Science CS Department: Highlights The UTD CS dept started as a small program within the Mathematical Sciences in the 70s One of the largest CS dept’s in the US today 55 faculty members 120+ Research and Teaching Assistants 15 Staff members including 4 Tech. Support Students (130 Ph.D MS +720 BS) Full range of programs in CS, SE and TE: -- BS, MS and Ph.D.

41 ________________________________________________________________________ Department of Computer Science Jonsson School of Engineering and Computer Science CS Department: Research Centers/Institutes  Cyber Security and Emergency Preparedness Institute (DFEPI) (Exec. Director: Dr. Doug Harris) - Cyber Security Research Center (Director: Dr. Thuraisingham)  Embedded Software Center (ESC) (Director: Dr. Farokh Bastani)  Human Language Technology Research Institute (HLTRI) (Director: Dr. Sanda Harabagiu)

42 ________________________________________________________________________ Department of Computer Science Jonsson School of Engineering and Computer Science CS Department: Accomplishments Ranked 29 th in UC Irvine’s publications ranking of CS graduate programs Ranked 24 th worldwide in UC Irvine’s publications ranking of SE graduate programs 8 of our faculty hold Young Investigator awards Top 5 producer of CS degrees Placed 14 th worldwide in ACM Programming Competition (just behind MIT & CalTech in US)

43 ________________________________________________________________________ Department of Computer Science Jonsson School of Engineering and Computer Science CS Department: Distinguished Faculty Over 55 memberships on editorial boards of computer science journals Research expenditure over $16 million in last two years Published 250+ papers last year Involved in numerous leading technical conferences as conference chairs or program committee chairs/members

Questions & Answers

Thanks for attending! Dr. Jeyakesavan Veerasamy