Air France KLM modernizes their payments service, SpringOne 2020 talk – Notebook

Air France KLM modernized their payments service recently, EPASS. This is a 12 year old system that provides the backend for processing purchasing airline tickets (and other things, I guess) from numerous front-ends: the web, mobile app, and social apps as well. The system was difficult to scale, it required manually adding new servers and had a long development cycle. As more and more people want to interact with Air France KLM through software (phones, online, in WhatsApp, or whatever other “channel”), they want to be able to evolve their software quickly. They want to use software as a core innovation tool for improving customer experience and, thus, business. So, here we see one of their first experiences modernizing their backend and transforming how they do software.

Talk presented by Oya Ünlü Duygulu and Patrick Zijlstra.

Highlights

-Rick in intro: transformed payments platform in 6 weeks.

– [Corporate vision] is to provide good, “our purpose as an airline group is create memorable experiences for our passengers. 

– So, they want to (1.) focus on customer centricity, (2.) innovation, and, (3.) efficiencies in our processes.

– “Digital” as the primary channel is on the rise. People want to interact through apps and such. So, KLM needs to meet the customers there… “As an airline, we want to be where our passengers are” (~3:00)

– Some examples of digital features: “‘About 10 percent of the ideas actually end up on the market. A recent example of this is the hand baggage check in the KLM app. Through augmented reality travelers can see whether their hand luggage meets the set dimensions. This function went live last month. ‘ Six months ago, a 3D rendering of the business class seats was also shown when checking in online. ‘This with the idea of ​​stimulating the sale of these chairs.”

– For example, listening and interacting with customers in social media [something I’ve done many times – it’s great to chat with someone (or a bot?) in WhatsApps, Twitter, etc. instead of a phone call]. (~3:40) Social media is now “our closest connection to passengers.” And in China: “For instance, Chinese travelers rely immensely on mobile devices. How can these personal devices be used for authentication – identity management, payment etc. to streamline the journey wherever possible? In China, the whole landscape is different, and we need to ensure we aren’t relying overtly on drawing customers only to our touch-points.” 

– (~3:10) merged together KLM and Air France backends to get less complexity in the back-end and a unified experience in the front-end for customers. Social media is now “our closest connection to passengers.”

– 350 agile teams. Using Scrun, Kanban. SAFe. See SAFe case study from ~2018, and some trip notes from someone at ACM Agile.

– The use SAFe release trains (which they call “release planes”), mapped to customer value journeys, e.g., sales, paid products, or airport.

– In the digital department, they have about than 50 product teams.

– Planning every three months, come together get a roadmap from the business, and all the teams plan together. Then they start sprinting bi-weekly.

– Also shared services and practices team.

– Their /goals: 

– (~5:50) “We are designing our products focused on time to market, innovation, robustness, and security.”

– Focusing on getting CI/CD in place.

– Also, reducing complexity and speeding up business value [realization], so we are moving towards a microservices architecture.

– [Business stuff:] EPASS handles payments from many places, created 12 years ago. Wanted to modernize [not sure why]. They worked with Pivotal Labs on modernization for a six wee project.

– EPASS app – made 12 years ago, handles about 37,000 payments transactions per day. Takes care of all online revenue.

– Six week engagement with Pivotal Labs. This brought expertiese from the outside, combined with their existing skills.

– “Six weeks is very ambitious for such a project, but getting this expertise from Pivotal and their dedication we made a success story at the end.”

– Modernization road-map for EPASS.

– We want to speed up with release cycles, which was then one month. [Move to single piece work-flow: whenever a user story is ready, then it can go live.] 

– In six weeks, all the could focus on was transforming the app and moving it to the new platform [PCF]. But, they could also modernize their skills by adding in TDD and pair programming.

– Switches to Patrick.

– (~10:55) – they go over their way of working. 

– Inception to set expectations. Outception to look back at what was achieved. Some blocker removal meetings. And the usual agile meetings.

– Two teams: one does modernization, the other delivers business features.

– Worked in one week iterations.

– Doing pair programming. “We noticed that this really increases the code quality that we deliver.”

– (~12:30) EPASS architecture. Was hosted on bare-metal Tomcat server. To scale, had to add new server and put EPASS software on it. This was becoming a hassle and fixing that was a motivation to move to VMware Tanzu.

– (~14:00) new architecture – five different components. Three in Tanzu Application Service.

– After, the majority of things were put in VMware Tanzu…

– [Picked some small things at first to test stuff out, hardcoded secrets but later fixed that – used CredHub – in long term will move to Vault.]

– (~16:00) Used Spring Boot, adding health check [this is good to highlight, that it gets instrumented/observable “for free”].

– “It was invigorating working to work with the  Pivotal experts and now there’s more confidence in the team to continue.”

– Used Bamboo, added in automation stuff for deployment…

– Problems: networking problems

– Benefits: response times improved by 10%; “all the power for scaling is within the product team itself” instead of having to work with other groups, file tickets, etc. Also, time to patch is within 72 hours (3 days).

– (~21:08) “The experience was very positive. It was invigorating to work with the Pivotal experts. And, now there’s more confidence within the team to continue to improve the application.”

– The projects have been finished for a few months. No more components in bare-metal Tomcat.

– “From the organization side, there is no more fear of big changes. If such an old application as EPASS can transform, then it’s possible for any application.”

– “So more and more and more applications will be moving to TAS [Tanzu Application Service].”

Manager, Heal Thyself! 

The panel I put together and moderated for this year’s SpringOne. It turned out really well. Here’s the abstract:

When you’re trying to improve how your organization does software, how do you change what managers and executives do? We hear a lot about how developers and operators change, the composition of product teams, and always about Kubernetes. But there’s very little conversation about transforming management. This panel of managers will discuss what managers’ and executives’ roles new and old look like, managing managers, and how individual managers can manage their careers when their role changes.

The panelests: Neville George, Manager at Comcast; Jon Osborn, IT Executive at Bell Tracy, Ltd.; Jana Werner, Head of Transformation at Tesco Bank

Questions for a panel about managing managers

A moderated a panel about managing managers during digital transformation stuff (organizations getting better at software, doing the DevOps, etc.). Here’s my vision for the panel and the questions we churned over. We didn’t directly answer all of them, of course. The panel was great! The recording should be up soon (it says September 10th 2020).

The idea/point/premise of the panel

In larger organizations, there are layers of managers, in a good way: teams aggregate to a manager, that layer of manager aggregates to another, then somewhere there’s executives, and, I don’t know, the mythical shareholder. Everyone has a boss. I want to discuss what it’s like to be the boss of all those managers and help them transform into all the existing, new fangled agile and digital transformation stuff. Most of the discussion I encounter is about individual staff and the product teams (those working on software or running it), but I don’t hear much about the management structures above those teams. Also, it’d be interesting to talk a little about what exactly things like “servant leadership” mean and how one manages their career (gets promotions, more compensation, etc.) when they’ve moved from being The Boss to a servant (to be tongue in cheek about it).

Questions

  1. We’ve heard the notion of servant leadership, which sounds, you know, helpful. Can you give me an example of what that looks like though, like an actual one that happened?
  2. I was watching a webinar that Jana did recently on her white paper. In the Q&A, they asked attendees something along the lines of “do you ever think of your organization’s vision and strategy, does it ever determine what you work on and how?” As I recall, almost zero percent of people responded yes. This seems like a critical tool for managers to use if they’re setting up autonomous teams that need to make decisions on their own – they need to know the principals, the goals. How should managers be moving beyond facile vision and strategy?
  3. For years, I’ve heard about “the frozen middle,” managers who don’t want to change despite the urging and permission of executives (“above” them) and enthusem of staff (“below”). Is this cliche real? If so, what causes that frozen-ness?
  4. (Following on from that), when you’re managing managers, what are you doing in this new, agile, world? Are you a servant to the servants?
  5. There are occasionally “accidental managers” who sort of ended up there. But most of them have been pursing a career of going “up” the meatware stack. They want to grow their career, which usually means responsibilities, the glory and power that goes with it, and the rewards. So, if you’re a servant to people below you, how do you end up managing your career?
  6. As you push responsibility down to teams, what are safety nets you put in place as they figure it out?
  7. What are some the first things you delegate?

Last chance to get $300 off SpringOne Platform

Next week is my company’s big conference, August 1st to 4th in Las Vegas. The agenda looks amazing, and it’s packed with good speakers. My primary interest, as always, is learning how orginizations are moving from older ways of doing software to newer, better ways. There’s a great line up of “manager types” talking about exactly that. There’s also endless coding talk, don’t worry.

Check out our recommendations for the conference in a recent Pivotal Conversations episode:

I’m going to try to interview people for the Lords of Computing podcast here and there as I find them. My lack of any real planning might ensnarl that, of course. On the other hand, locking people’s schedule down for a speaking at conference is next to impossible. That’s my excuse at least.

If you’re interested and haven’t registered yet, you can still (I assume!) use the code pivotal-cote-300 to get $300 off registration.

Hopefully I’ll see you there!