Tell me about yourself and your job role:

Sudeep Chatterjee: I am a Senior Technology Manager who has more than 19 years’ experience with top-tier investment banks, fintech and consulting firms; managing testing globally for enterprise-wide change programmes. I am currently consulting as Head of Testing at Bank of America Merrill Lynch within FICC – Global FX Technology group. Prior to the Bank of America Merrill Lynch, I have worked as Head of Testing with Lombard Risk, Barclays, UBS, GE and Accenture, while primarily focusing on building high performing multi-disciplinary testing teams and delivering testing for complex technology-driven business transformation initiatives.

Niranjalee Rajaratne: I am an IT professional with over 12 years’ experience who has worked in various verticals such as telecom, financial services, publishing and e-commerce. Currently, I am the head of quality assurance at Third Bridge. Third Bridge is a leading independent financial research company that provides private equity firms, hedge funds and strategy consultants with the information they need to make an informed decision about investment opportunities.

What is the most complex system architecture you have managed?

Sudeep Chatterjee: I have managed testing of many complex system architectures starting from PoS systems for one of UK’s largest retail company, complex integrated voice recognition systems for one of UK’s largest telecommunication company, complex banking systems built over quants and big data solutions and recently work on the high frequency low latency electronic trading platform for FX.

Niranjalee Rajaratne: One that was developed on unstable grounds where the system was built and managed in an ad-hoc manner. There was a poor team structure in place and there wasn’t enough skills and knowledge in the teams. Also, the business knowledge was not built into the system properly. This created uncertainty when it came to software testing and delivery. It became difficult to achieve quality and to take proactive measurements to risk management.

Are your change management processes modernised and agile friendly?

Sudeep Chatterjee: Change management process is agile and focuses on ensuring SDLC is built over robust processes, which allows for faster delivery with high quality.

Niranjalee Rajaratne: Yes, up to some extent. We keep evolving. We try to iterate the change management process so that it enables us to achieve the business needs faster, better and within the budget. As an outcome, we try to do early engagements with business units, improve communication to keep it consistent and get the buy-in from senior management as early as possible. The team is given autonomy and time to try new things and adapt to change as new and brilliant ideas come to light.

How do you stay ahead of the competition when it comes to mobile systems, cloud computing and APIs?

Sudeep Chatterjee: Adopting a modern digital transformation strategy is part of the technology roadmap, which includes best-in-class delivery for mobile platforms and cloud solutions for end users.

Niranjalee Rajaratne: Research and assessments help us to continuously refine and adopt the right level of technology to achieve business objectives. We assess our current workflows and software to identify gaps in efficiency that technology can bridge.

Can you give me an example of a CI/CD failure you’ve experienced, and what you learned from it?

Sudeep Chatterjee: CI/CD has gone wrong mostly in my experience, especially when dev teams are used to implementing plans and runbooks for deployments in non-production and production environments. When teams start using CI/CD tools without proper training there are times when an issue comes up, particularly around configuration setups.

Niranjalee Rajaratne: I can give many, such as the wrong selection of CI servers, inefficient build infrastructure, uncorded and unstructured pipeline, incompetent skill set and lack of management support. It resulted in delivering business value to the end user albeit on broken builds. CDs cannot be performed and organisation cannot be productive with an inconsistent CI. To make CI/CD work, it was important that the business and technology management understood the value of an efficient build pipeline, and support it without seeing the effort as pure technical work.

How has DevOps changed your way of working?

Sudeep Chatterjee: DevOps has helped organisations to reduce the time to deliver software to production with continuous testing embedded in the process. It has also strengthened testing frameworks like behaviour driven development.

Niranjalee Rajaratne: DevOps improved infrastructure and created efficient build pipelines to help get new products and services to the users quickly. It created a shift in testing and quality assurance by enabling automation of repetitive tasks such as regression and smoke testing. The development teams were empowered to follow test first development practices such as TDD and BDD, helping them towards an efficient build pipeline, and to receive fast feedback.

What DevOps tools do you use?

Sudeep Chatterjee: Jenkins, Git, BitBucket, Maven and Ansible.

Niranjalee Rajaratne: Github, Ansible, Docker, Jenkins, Elasticsearch, Monit are few.

Have you faced any bad experiences when implementing DevOps?

Sudeep Chatterjee: Implementing DevOps without organisational culture change can cause conflicts between teams. For successful DevOps teams, it is important the change management process complements the ethos, and all team members from BA, developers, architect, environment management, QA and application support must understand their role in the DevOps world.

Niranjalee Rajaratne: We certainly did have challenges and we still have some to this day, but I would not say they were bad experiences. Rather they were learning opportunities for the team to perfect the delivery pipeline.

How will DevOps continue to improve software delivery?

Sudeep Chatterjee: DevOps will continue to improve and will be ‘the norm’ for all software delivery. Dev and QA teams will learn to work with DevOps tools including the continuous integration and continuous delivery de-facto standard.

Niranjalee Rajaratne: It will further support the team’s autonomy to be self-organised by removing dependencies, further breaking silos. It will continue to improve inter-team collaboration, empowering people to learn new skills and share knowledge. It will bring users and technology teams closer to achieve a symbiotic relationship. If iteratively improved, this will enable organisations to take a paradigm shift in how they deliver software.

Do you believe manual testing will come to a close because of automation and DevOps?

Sudeep Chatterjee: There will always be a requirement for manual exploratory testing by domain experts though this may not just be a QA professional but can be anyone in the team with strong domain knowledge like BA, product owner, or users.

Niranjalee Rajaratne: The purpose of manual testing and the value it brings cannot be fully replaced by automated testing. Whilst automation testing is an integral part of DevOps, manual testing should continue to function in order to help achieve optimum levels of quality in software.

Anything else you would like to add?

Sudeep Chatterjee: One of the challenges in the agile and DevOps world will be how organisations measure product quality and compensate its employees for performance management. You must make sure the HR process matures to be to able to provide compensation to the entire team, or you will have to still rely on individual feedbacks and do performance benchmarking between team members.

The post How to keep on top of testing techniques & tools appeared first on DevOps Online North America.

DeployHub outperformed competing solutions such as Microsoft Visual Studio Team Services and Puppet.

The Gartner report indicated DeployHub was made “attractive to new teams looking to get started with ARA”, because of its affordable price, as well as its “strong integrations” with Ansible and Jenkins.

Tracy Ragan, CEO at DeployHub, stated: “We have worked hard to deliver an ARA solution that is highly functional and low cost, allowing development teams to get their hands on a continuous deployment solution early in the process.

“We are particularly happy that Gartner acknowledged our integration work around both Jenkins and Ansible.”

Written from press release by Leah Alger

The post DeployHub ‘outperforms’ Microsoft and Puppet appeared first on DevOps Online North America.

Even with a helping hand from developers, testers, deploys, monitors, log, configuration management and collaboration platforms, the online gaming company’s road to DevOps wasn’t straightforward. “Our road to DevOps consisted of monthly releases, downtime during release, costly meetings between teams, communication issues and late integration issues,” said Sztupak.

In a bid for Gamesys to change its path another route needed to be found. “We needed to find a way to prevent the following issues: we needed to split up monolith, create a platform that allows people to create microservices easily, deploy those microservices, automate processes and automate what’s on the Dev’ side,” he added.

Before the change, backend and frontend teams were separated, teams were effectively going at different speeds and frontend and backend teams were merged into verticals.

“A platform team needed to be built up from members of different teams. We set up to drive the move to microservices, gave ourselves more leeway in assessing new technologies, created Dropwizard (a common platform) and found a way of specifying our APIs,” revealed Sztupak.

After 5 months and 3 weeks they finally found solutions. Legacy-in-a-box removed the convoluted build and deployment process, and made it easy to make changes to the legacy monolith.

“To ensure that the following was achieved, teams started to create microservices even though there was no way of deploying them, containers were introduced, GoCD was used as a framework, and everything was built automatically in git,” he added, concluding that Ansible, Docker and GoCD are the backbone of DevOps technologies.

Written by Leah Alger

The post Gamesys road to DevOps appeared first on DevOps Online North America.

“By 2020, 50% of global enterprises will have implemented at least one application release automation solution, up from less than 10% today,” according to Gartner’s new ARA Magic Quadrant report.

Application release automation and DevOps

This is the first Magic Quadrant for Application Release Automation that Gartner has ever produced – a clear indication that ARA tools have a big role to play as DevOps teams strive for a continuous delivery pipeline that rapidly turns out small releases in a stable and secure fashion.

The ARA market is just six years old, but it has already grown from US$20 million in 2010 up to US$219.7 million last year, according to the report. And with enterprise adoption poised to grow by a factor of five in the next three to four years, it looks like we are at the beginning of significant investment by enterprise IT teams. Companies trying to take DevOps to the next level are looking to ARA tools to deliver.

It may be a relatively new market, but there are a number of players already jockeying for position.

There’s a strict criterion for inclusion in Gartner’s Magic Quadrant. Only vendors offering a comprehensive ARA solution that includes automation, environment modelling and release coordination functionalities were considered. All the companies named are also generating at least US$1 million in annual revenues solely from ARA products licensed to at least 20 paying, production-level customers.

Leaders in the ARA field

While big names like Microsoft and Puppet are described as niche players, Gartner has singled out Automic, CA Technologies, Electric Cloud, IBM, and XebiaLabs as leaders in the ARA field.

“We believe that being placed in the leader’s quadrant is a clear recognition of our product innovation and our focus on the needs of enterprise IT teams,” said Derek Langone, CEO of XebiaLabs. “It is a testament to the tremendous value our global enterprise customers have been able to create in a short period of time, as they use XebiaLabs to power the software delivery of mission critical applications, from flying satellites to space to managing transactions worth billions of dollars.”

One big reason a wide variety of industries are turning to the top ARA tools, from financial services to retail, to transportation and beyond, is the ease and speed with which they can be implemented and can scale for large organisations. This simplicity represents a significant improvement over the alternative of building custom programs and scripts to link other tools in the pipeline.

IT teams have made costly and time-consuming mistakes by trying to retrofit continuous integration tools like Jenkins and Bamboo, or provisioning tools such as Chef and Ansible. Application release automation is designed to be a central hub where companies can plan and orchestrate their software release pipeline and ensure continuous and reliable deployments.

Automation is the future

Gartner’s report mentions automation, environment modelling and release coordination as the key evaluation requirements that enterprises are looking at. ARA tools help teams to dispense with inconsistent manual methods at every stage of the product life cycle. They’re enabling businesses to manage and co-ordinate an increasingly complex set of tools, processes, and systems.

Accessibility and scalability are important too. Businesses need to be able to visualise and comprehend the progress and current status of all active releases. They also need the freedom to add new tools and technologies and integrate them into the pipeline. Easy integration with the latest technologies allows customers to retain maximum flexibility and business agility.

There’s also demand for a platform that’s accessible for management and provides a rich set of metrics. Managers need to maintain control and visibility, without hampering the autonomous decision-making that enables DevOps teams to innovate. Some orchestration tools are just too complex and technical; if you can’t extract actionable insights, all of your metrics are going to waste. And if you aren’t gathering all the right data in the first place, there’s no insight to be found.

Gartner’s ARA Magic Quadrant is defining this new space. The firm is confident that ARA tools are set to take centre stage as enterprises try to expand on the agility gains they’re made from DevOps and other automation initiatives. Companies have already dipped a toe in the water, now they’re ready to dive in with application release automation and make the promise of continuous delivery a reality.

Written by Simon Hill. Edited for web by Cecilia Rehn.

The post Gartner’s new ARA Magic Quadrant highlights future for DevOps appeared first on DevOps Online North America.

As companies work smarter and introduce new technology, they’re increasingly looking for people who can enhance their technical capacities and drive development efficiencies in the process.

IT recruitment agency Amsource Technology has compiled some advice on the skills needed to succeed in the world of DevOps.

Experience with configuration management tools

Candidates who can confidently use configuration management tools such as Chef, Ansible and Puppet will likely have an edge over those who cannot. These tools can help to support and optimise software deployment, and there are plenty of options to choose from. What’s most important is the skill level of the DevOps professional wielding the tool, so whichever you choose, make sure you know it well.

Soft skills

Soft skills can be as important as hard skills, if not more so, for employers. A 2013 Vanson Bourne survey found soft skills were the most in-demand attributes in a DevOps candidate, showing the need for professionals to be well-rounded. Clear communication, collaboration and flexibility are all desired by employers in the field, so jobseekers would pay to focus on these skills alongside their more practical, technical knowledge.

OpenStack knowledge

According to a Dice salary survey, DevOps workers with OpenStack skills were amongst the top five highest paid employees. The cloud computing programme has set itself apart from the rest of the industry with its open source model and collaborative development process. It helps engineers to work quickly and maximises existing infrastructure, allowing for autonomy and innovation.

Learn programming languages

Knowledge of programming languages is key for a variety of tech roles, and there’s no exception for DevOps. Some of the more popular languages in this sphere include Ruby, Java, Python, Shell and Perl, but there are plenty on the market that offer something slightly different. Ideally a DevOps professional will have strong working knowledge of more than one language, so if you’re looking to transition into this field or expand your career development opportunities, train up in a few.

Infrastructure insight

DevOps engineers in particular should have a working understanding of datacentre-based and cloud infrastructure components. This includes aspects such as how software is racked in networked, methods for load balancing, running virtual machines and networking basics such as layers, networking in cloud and access between applications.

Infographic

The team at Amsource Technology have compiled a helpful infographic summarising the skills information.

Edited from press release by Cecilia Rehn.

The post Advice on key DevOps skills to brush up on right now appeared first on DevOps Online North America.