If you are battling to keep up with trends in the software industry, by continuously releasing enterprise products, adding more functionalities and improving user experiences, then someone needs to focus on who is protecting the code. Somebody needs to look into the process and methodology of how these code quickly gets built and rolled out and still protect the code.
Welcome to the world of coding, coders and a whole million lines of code.
“Oh I am not a monster, Its just that I am ahead of the curve” ~ Joker in Bat Man. 🙂
Most often so, that which is usually done by operations team within large enterprises, today, a term called “DevOps” has set its large foot. Code is not only protected but deployments are continuously happening. What is being seen is, there are no huge benefits of two weeks roll out. According to an independent study, if these two weeks of roll out is not carefully planned, you are bound to constraints that will extend your life cycle of the project in general. The end results that you would have attained within say X amount of time would cost you X+N amount of time. So check the plans. A small tab sheet for the devops is helpful anyways.
Are you having that tab?
When Georgi from this startup company came to me asking for help in safeguarding the huge set of code base, I didn’t realize that the damage has already been done. There are things that these guys have to do and have to do real quick, before it goes out of hand for a solid spin. You see, Georgi was smart enough to approach with his little doubts on issues he thought may pop up. It was a wise move. So here is the use case and the solution provided. Total cost involved, number of people, time lines and methodology adopted for implementation and delivery has been highlighted. Read on.
Scenario:-
- 16 Developers with 7 developers additional remote.
- 5 Testers one onsite and rest remote.
- Four different components; Front End UI, Event brokers/messaging layer as a transport layer to front end and back end platform, the database access layer and finally the platform by itself. 32 Integration check points including connectors to search repos, analytics repos and posting data to other 30 internal and external systems.
- Two to three weeks delivery cycle. Technically something was rolled out every two weeks as per Georgi who was managing all these activities and was the direct report.
- Back up servers not timely but on demand.
- SVN and Git local repos. However no synched up code as there were multiple developers checking out code from the same branch and same class files or code.
- The development environment was using Java, Javascripting, nodejs, rabbitmq, MySQL, Solr, Lucene, zookeeper, spark and spring boot and hibernate, redis, nginx and tomcat. Containerization was done using CoreOS and docker. Eventually they were planning to move towards one and consolidate it. I am working on evaluating CoreOS and docker for them and seeing which suits their environment.
- Currently there performance is being assessed through custom built performance tools and network access and speeds are tested through common tools and simulated users originating from outside the firewalls.
- Build tool is Maven combined with Gradle , Chef recipes, bamboo and jenkins interacting through hooks to Git. Also there is Bug Reporting tool for which integrating is asked for. SVN is standing alone with all javascripts being posted there. Reason was, We began version control and management with SVN.
- Cloud enabled.
Solution Visual
