paint-brush
Modern Cycle Development Techniques, DevOps and Agile for Retailersby@DataArt

Modern Cycle Development Techniques, DevOps and Agile for Retailers

by DataArtAugust 1st, 2017
Read on Terminal Reader
Read this story w/o Javascript
tldt arrow

Too Long; Didn't Read

On July 12, 2017, Anton Krasikov, Senior Technology Advisor at DataArt, spoke at the <em>What Retailers Need to Know about Microservices, Containers &amp; Data Services</em> Summit, which was organized by Mesosphere and held in London.

Companies Mentioned

Mention Thumbnail
Mention Thumbnail
featured image - Modern Cycle Development Techniques, DevOps and Agile for Retailers
DataArt HackerNoon profile picture

On July 12, 2017, Anton Krasikov, Senior Technology Advisor at DataArt, spoke at the What Retailers Need to Know about Microservices, Containers & Data Services Summit, which was organized by Mesosphere and held in London.

The summit gathered together CIOs, CTOs and enterprise architects of large retailer firms with the objective of offering first-hand insights on modern development cycle techniques and sharing best practices on how retailers can deal with the challenges of delivering personalized services to consumers in real-time.

One of the topics was Leveraging DevOps in Retail, presented by Anton Krasikov, Senior Technology Advisor at DataArt. Anton has 15 years of experience in IT, including 4 years working on projects (Java, DevOps).

Krasikov used practical case studies to outline the gradual evolution of DevOps and its application in projects for clients like Ocado and other retail companies. He provided an overview of how companies moved from physical infrastructure to virtual machines, and then to containers. Anton also highlighted current challenges in managing containers and offered potential container orchestration solutions and an infrastructure management approach using DC/OS.

Anton first introduced DevOps and discussed its use in retail. He emphasized that DevOps aligns very well with a microservices approach..

He went on to introduce the concept of DevOps, and the principal idea of bringing development and operations together, with developers understanding the infrastructure and operations understanding the code.

The DevOps approach is characterized by the following:

  • Continuous integration — when a code changes a new package is created, tests are run and it is immediately evident if something goes wrong, making it possible to react accordingly.
  • Continuous delivery — after your package is continuously integrated, the code can go straight to production, ideally, or to further environments. Any changes made to the code are tested, and then you have a single immutable package that could be promoted from one environment to another, up to production. Continuous Delivery reduces the level of human interaction to avoid risks such as forgetting to update the database or a configuration file.
  • One of the core concepts of DevOps is Infrastructure-as-a-Code, in which you are basically defining the whole infrastructure in a simple manageable text format, and you keep it in a version control system like Git.

The DevOps concept helps respond to rapid changes. It helps teams work on various micro services together and is a very good fit for Agile. It is no longer a buzzword, but it is still a very good choice for implementing an iterative approach.

When you apply DevOps in your organization, it’s a big cultural shift for people to understand that infrastructure can and should be flexible.

Anton also shared his vision of the future of DevOps. It is a rapidly developing concept, as is microservices, and there is a huge debate on container standardization. There is an Open Container Initiative, which is largely based on Docker, but there are still many companies trying to do their own versions, including Google and Amazon. Container Orchestration tools and methods are also being actively developed and we can expect major improvements in this area.

Another topic was microservices in retail. Terry Dhariwal, a Solutions Architect at Pythian, discussed the anatomy of microservices architecture and the key factors of successful implementation. Terry also explained why microservices are relevant to retail.

He delivered best practice examples, which included Tesco, British Gas, Dixons Carphone Warehouse, Sky and Experian.

Terry emphasized that one of the key aspects of microservices is making things highly scalable across the whole stack. The idea of microservices is to split all product functionality into small components, each with a single responsibility. For example, Tesco has microservices for basket session or shopping cart. Home delivery is also a microservice. The idea is to componentize in order to avoid creating a monolithic application, as this stops innovation and causes disruption. The approach “one architecture — one application — one step” no longer works in today’s time to market, as it stops a team from innovating quickly.

The idea of Microservices is that the team of developers, DBAs and DevOps works together and understands the application not just for development, but for production too.

Microservices are very important for retail as it has many moving parts, including distribution centers, customer service, basket sessions, and price changes in a very dynamic environment. If a retailer doesn’t have a microservices strategy, they risk getting stuck when it comes to innovating.

A key factor in successful microservices is making the entire stack elastic and scalable, highly available, with consistently high performance and the agility to quickly innovate. The infrastructure also needs to support these characteristics. For this you need to go to cloud, where things are more standardized and secure.

Moreover, when it comes to microservices automation is key. When you have to deal with many services and applications, without automation you are bound to cause disruption when deploying new features. And the database is very important, as is leveraging DevOps.

After the presentation there was a brief demonstration by Mesosphere’s Richard Lawrence on how to build a retail analytics big data stack in minutes to utilize real time customer data and profile for product recommendations.

Richard showed the features of DC/OS beyond a scheduler (like Marathon) and demonstrated the benefits of Mesos as a 2 tier scheduler, fine grained resource allocation and Frameworks that understand the application it is supporting.

About the speakers

Anton Krasikov

Anton Krasikov is a Senior Technology Advisor at DataArt who has been working in the IT industry for more than 12 years, 10 of which have been with DataArt. He has participated as a developer, team leader and solution architect in various projects across the retail, telecom, travel and financial services industries. Anton has a keen interest in Big Data and DevOps technologies, and enjoys meeting the challenges of building highly available and resilient IT systems. He is a regular speaker at DataArt “IT talk” and “JavaDay” conferences, speaking on subjects ranging from DevOps innovations to client communications. Anton has an MS in Computer Science from Voronezh State University. [email protected]

Terry Dhariwal

Terry is a Solutions Architect at Pythian, with a background in real-time analytics and personalization, dealing with high volume and high throughput data processing systems for the retail sector. He supports customers to deliver enterprise applications that address the performance, scale, availability, and flexibility needs for modern day mission-critical applications. These have included Tesco, Dixons Carphone Warehouse, British Gas, Experian and Sky.

Originally published at http://blog.dataart.com/modern-cycle-development-techniques-devops-and-agile-for-retailers/ on July 12, 2017.