back to overview
15
.
05
.
2020

Ready for a virtual journey to China?

A short story of how Swiss Re's Coloride became 瑞再车睿驾
András Kelemen
CTO Automotive & Mobility Solutions at Swiss Re, Co-CEO at Movingdots GmbH
Andreas Oellien
Head of Operations & Application Management
Martin Webermann
Head of App Development
Dennis Wiemann
Head of Services & Integration

China is a huge market that many enterprises are striving to tap into. Still, there are a few technological barriers to overcome which might seem discouraging on the surface. Well, we have just converted our car telematics solution for the Chinese market, so read on if you would like to benefit from our experience. It is an entirely digital product consisting of an IoT platform, along with corresponding apps for iOS and Android called Coloride. The platform processes GPS and Accelerometer values from the app and enriches them with cartography information from a map provider. This information is then analyzed in order to calculate the driving risk for motor insurance. The technical complexities in China affected our entire solution, from Cloud hosting on Azure to specially encoded GPS coordinates as well as the unavailability of Google libraries for Android. Let us elaborate on these aspects in more detail, step-by-step.

Although most of the engineering work was done in Europe, we should also mention the local Swiss Re colleagues in Beijing who helped us with adapting to local specifications. Since each digital product is different, what we did might not be applicable for every company. Still, there might be elements which are quite useful.

Team setup

Agile organization

Our product development team is organized into smaller groups that all follow the Scrum methodology. App Development, Platform Development, Application Management, Data Analysists: All of these separate groups, working from different locations, needed to be aligned and synchronised. What we usually do when we start working on larger features, such as the "China go-live" project, is setup a feature team. And that is exactly what we did in this case. It was extremely helpful to have a strong feature lead, a well-experienced project manager, who was located in China and who had previously worked in Europe. She was an invaluable help in bridging our European teams with the Chinese culture. She made sure to fill our backlogs and sprints and gave valuable insight to the team regarding the Chinese specifics. Having strong support from someone who was familiar with the Chinese market and language turned out to be a great asset.

Legal support

From day one, we involved data protection experts and legal counsels. Our experts from Switzerland explained and covered all aspects of our complex products in relation to data protection and data flows. They also collaborated with the legal team in China in order to identify any roadblocks and to implement any necessary adaptions. Last but not least, they supported us replacing Coloride with a more appealing Chinese name, 瑞再车睿驾. Team members with different backgrounds and who come from different cultures all collaborated to successfully achieve the same goal

Adapting the mobile app

Coordinate systems

Due to national security regulations in China, it was necessary to implement a different, scrambled coordinate system (called GCJ-02) that required a special authorization from the administrative department of surveying and mapping under the State Council. To enable 瑞再车睿驾 (incl. all underlying map and location services), we extended our trip processing to support another coordinate system at the same time. Every trip that gets recorded inside China runs through a specific trip processing procedure with suited map services. Afterwards, the app decides which trip should be displayed based on the current location of the smartphone

Distribution channels

The distribution of Android apps is far more fragmented in China compared to other regions in the world. Since the Google Play Store is not available in China, Android apps are being provided through more than 400 different app stores. As a result, the efforts for publishing Apps increases with the number of Android stores. For the time being, 瑞再车睿驾 is being published to three of the biggest app stores available: Huawei, Oppo & Vivo.

Non-Availability of Google Play Services

Unlike most regions in the world, Chinese Android smartphones are not delivered with pre-installed Google Play Services framework. This framework provides several helpful functionalities to build and further improve a telematics app. The Google Location and Activity Recognition API is used in Coloride as one of the existing triggers/mechanisms to detect when users start or end a trip. As part of our entry into the Chinese market, we had to replace these functionalities in 瑞再车睿驾 with similar frameworks and different workflows.

Deploying to Microsoft Azure in China

The first milestone was set and the decision to launch our telematics product in the Chinese market was taken. There are two strong reasons as to why we needed to deploy the infrastructure on site in China. On one hand, it was because the latency in API calls from China would render the solution unusable if the apps were to connect to a data center in Europe. Furthermore, due to the regulations in China, it was necessary to keep and process all the data within the country. To comply with these regulations, global technology providers do not offer their cloud services directly but select a local provider which operates their cloud technology stack independently.

Azure Services

The MS Azure Tech Stack operated in China has been lagging behind the ones operated elsewhere. Not all necessary Azure Services were available or, if they were, only in older versions. The deployment of some of the not yet available services, e.g., Azure B2C AD had already been announced and the launch took place before we started our deployments. While others were still unavailable, such as Data Bricks, Application Insights and Log Analytics. The unavailable Data Bricks service was not really make-or-break to begin with since our core platform was able to run without it. Nevertheless, not having Application Insights and Log analytics in place was an issue, because they are extremely vital for our monitoring and alerting and we couldn’t guarantee operations without them. That’s why we considered an alternative. The easiest solution for us would have been to minimize our logging content and implement a hybrid solution where we would host Application Insights and Log Analytics in Hong Kong or Singapore while all the other services would be deployed on Mainland China. We started to clarify all the necessary legal requirements for the hybrid solution and did some test runs but, in the end, it wasn’t necessary to begin the implementation of this scenario. Fortunately for us, at the last moment, our direct contacts at Microsoft were able to speed up the internal processes and they announced the launch dates of the previously unavailable services. So, we abandoned the testing of the hybrid scenario and shortly afterwards successfully deployed our backend, including the logging services on Mainland China.

Local sign-in practices

When we first showed our app to the Chinese team, one of the first reactions was that we offer an unusual login procedure. Entering email addresses and passwords is not common in China. We also found this out while working on a project at the same time in Saudi Arabia, where people sign-in with a phone number. Thus, we had two target markets asking for the same adaption which was reason enough to get started on the project. Luckily, around the same time, a public preview was made available from Microsoft and allowed the setting up of a password-less sign-in feature as part of Active Directory B2C. Still, we had a few hurdles to overcome: The One-Time-Password solution only supported standardized, English messages to get sent to users, so we had to connect our own service with our own templating system in order to be able to  define messages, depending on the region. Although it was not easy, translation was an obstacle that we were eventually able to overcome. In conclusion, there are definitely less challenging ways than using a service in state of preview, but we finally managed to get to a production state that we could use to showcase our solution in China.

Conclusions

One of the most significant success factors was the interdisciplinarily organized work in our distributed team setup. Our main challenge was to find solutions for different cloud infrastructure particularities and other regional circumstances. With the involvement of local experts, we were able to find the right answers for every question. It was overall a very rewarding experience for all of us to see how the pieces of this big puzzle got into place step-by-step, and how Coloride became 瑞再车睿驾.


Share Article