Advertisment

A key component for faster delivery

author-image
Sunil Rajguru
New Update
devops

Dr Niladri Choudhuri is CEO Xellentro, which gives training project, program, and portfolio management. Here he talks about the relatively new concept of DevOps and how organizations should go in for it to optimize their IT operations.

Advertisment

DevOps is a relatively new concept. Which are the types of companies which have been the early adopters? Who are currently missing out and you would recommend implement it immediately.

There is no specific type of company for this. It is now 10 years that the DevOps name was founded by Patrick Debois. Previous to that people used to do it in bits and pieces without knowing it was part of DevOps Values. Today every organization needs IT to run their business. Everyone needs to deliver in a faster, stabler, happier and securer manner. Thus DevOps is required for everyone. It can be the existing organisations or even startups. The basic concept is to break the silos and work together for the same objective. This is true for every organization and not limited by any domain or geography or technology. But it is easier for a startup to implement, while it is a big task to change the bigger organizations. The bigger it is, more difficult it is. A lot of it is about thinking in a new way and taking the help of automation.

What are some of the main errors in the way DevOps is being implemented in India and what are the ways to rectify them?

Advertisment

The main challenge I see is that organisations are getting into DevOps by just using various DevOps tools that have come into the market, like, Chef, Puppet, Ansible, Selenium, etc. These are required to deliver faster and will only give advantage of automation. But they will not give the benefit that is expected. First of all the Value Stream Mapping has to be done on each type of changes and then one has to decide on the areas of improvement and thus the Toolchain. Proper CI/CD requires a holistic view and loosely coupled architecture.

Management needs to be clear in their vision. They have to be objective, have patience, invest in training, culture change and tools and other support for creating a learning environment to embrace DevOps properly. They are used to a way of working and there is a need to change the mindset of senior management. I feel they need more help and handholding than the developers as they are more adaptable. However not much help is available in this area and hence we are starting with an initiative DevOps for Executives bringing in a few experienced experts of DevOps to create a forum for leaders.

What about the size of the company? Is there any impediment in smaller companies implementing DevOps?

Advertisment

This is applicable to any organization irrespective of its size. Size is not the criteria to decide whether to do DevOps or not. The criteria is whether the product/service of the organization needs faster delivery. For example, a timesheet system does not need frequent changes while an ecommerce site does. The point to see here is whether the service or product needs faster delivery. It is easier for smaller organizations to implement DevOps. The only challenge that I see is that it might face difficulties in getting the fund that needs to be invested at the beginning. The result will start coming only after some investment in terms of fund, time and resources till the new process flow stabilizes. The State of DevOps Report 2019 shows that DevOps is able to get good results from organizations with 1-4 employees going up to 10000+ employees.

How can software engineers upgrade their skills to keep up with the world of DevOps?

From the individual perspective, they should first go through DevOps Certifications like DevOps Foundation, Continuous Delivery Architecture, DevSecOps Engineering, etc. from DevOps Institute, learn some of the tools of DevOps and after that practice to make it perfect. Organisations need to support these. They should also attend and even speak in conferences. From organization perspective, there should be Centres of Excellence created, Dojos or place of practice should be there where people can come and learn and experiment with a charter to prove or disprove a hypothesis. These should have infrastructure, tools, trainers/experts/coaches along with SMEs from business. People should be given time within their normal working hours to experiment and learn and not do things on their own time. This requires investment and organization should be ready to spend that. In long run, it will help and reap great results.

Advertisment

What about similar concepts like WinOps and Google’s Site Reliability Engineering? How do they fit with DevOps?

SREs are good compliments to DevOps. It is Developers managing Operations, fitting very well with the Principles of DevOps and the DevOps Values. It helps in better collaborative work and Sharing. SREs help in focusing on Major Incidents, Problem Management and thus making the applications stable and thus have less instability and high availability by improving MTTR. There will be various names like WinOps coming up. WinOps is about practicing DevOps with a Microsoft product centric view. The basic principles do not change.

How will technologies like AI and ML change DevOps?

AI and ML will make a huge difference in DevOps. One of the major areas is Monitoring. We would like to have Monitoring not only in the Production environment but also on Test and in more mature situation on Development environment. We can say something will work only when we have tested on Production/near Production environment with real users doing real transactions. AI and ML can help in the improving these areas to get it into self-healing systems and have a lot more pro-active monitoring and thus improving MTTR. This increases the availability of the system and thus more transactions can go through thus generating more revenue for business. Change Management will also be improved in the lifecycle with AI and ML and thus improve further improvement in availability and capability. With the use of cloud, AI and ML in DevOps, it is the best combination for any organization to scale their business.

devops
Advertisment