Developing Applications on the Cloud with Google Kubernetes Engine
Software Partner Inc.
System Solution Group 2
（Pictures, from left to right）
・Mr. Katsumi Kureishi
・Mr. Yasunori Watanabe
As a system integrator that works to meet clients’ needs around system development, infrastructure building, product development, and product retailing, Software Partner Inc. (hereafter Software Partner) has implemented and utilized Google Cloud for the System Integration (SI) of golfing-related application “Furutoru”.
Overview of Systems and Services
We developed a program that could film and play back golf swings named “Furutoru” at the request of a client.
There are 4 unique characteristics of this service as follows:
1. Simultaneous filming from multiple angles
2. Users can view videos through the internet on smartphones and personal devices
3. Frame-by-frame navigation and slow motion playback options
4. Allows users to easily impose a trajectory line over the video by gesturing on the touchscreen display
For features to distinguish it from the competition, it enables simultaneous filming from the front and side, and also allows for a trajectory line to be imposed, resulting in a system that enables users to check their form for in-depth analysis.
Our system is designed around the vision that it will be deployed to numerous golf practicing facilities in the Tokyo area, utilizing Cloud Armor for security and Cloud Load Balancing for dispersing processing loads. This allowed us to create a highly scalable system.
Issues and Goals Before Implementation
There was already a platform that processed on the end-user side, and since modifying that system would have taken time and money, they decided to develop a brand new system. They decided to go with our idea, and that’s how we decided to go ahead with this project.
However, in operating our service through the internet, we felt that there were serious challenges in the form of security and storage considerations, so we decided to use cloud technology.
In addition, in building our system to operate through the cloud, there were concerns on how we would securely operate things such as online inquiries, service management, and data stored.
Why We Chose Google Cloud and Cloud Ace
We began to compare options for cloud services after deciding on using cloud technology, and our experience at Cloud Ace’s Big Data & ML training (*1) in 2019 gave us some insight. There, we found out how Google Cloud products were easy to use, and we became interested in them as technicians. That’s why we started considering implementing Google Cloud.
We felt that Google’s “zero trust” policy was one step ahead of the competition. However, there were definitely areas where we lacked the technical knowledge to understand well. We still decided on them because of the IaC (Infrastructure as Code) that we had learned in Cloud Ace’s training, because it would prevent mismanagement of settings, and a place that unified controls. Further, the fact that we could leave the security settings to Cloud Booster (*2) was another factor in our decision to go with Cloud Ace. We felt confident in leaving our settings to Cloud Booster, because their members exhibited deep knowledge of security matters in a hearing session.
Additionally, on-premise solutions would require consideration towards security and peaks in system usage, but through the use of Google Cloud we were able to give minimal consideration towards security and scaling.
Even better, we could be flexible in shifting scales of operation, all without stopping services.
About Cloud Ace’s Services
Our timetable for this project was very tight at 2 months, and for a Google Cloud novice with limited time like us, Cloud Booster was a very helpful and reliable guide.
We are very grateful for the support and lectures about Google Cloud basics and billing, system environment, processing load, security and other topics.
They provided us with an environment where it was easy to ask questions on Slack. They responded quickly and thoroughly, hosting online meetings when necessary, and we were able to solve issues without taking too much time.
They also kindly responded to non-technical questions relating to operations, such as cost-saving measures.
Additionally, we also received support for preparation in using Google Cloud. Being amateurs in the use of Google Cloud we were very grateful, and were able to build an environment that balanced performance, cost, and security, for a system that was efficient.
Effects of Implementation
We would have had to down-scale our plans for cost and time concerns if we were to build infrastructure on-premise.
By combining the various cloud services of Google Cloud, we were able to go for a non-compromise solution.
Additionally, we are confident and satisfied with the managed services.
We were also able to lighten the maintenance load by going for a cloud service.
Furthermore, we were able to lighten our workload and have an unrushed development process due to the automatic verification and deployment functions, as well as code committing and modular builds.
Through this round of lightening the system, we would like to make it so that we can quickly and efficiently update it.
At this stage our program is bare-bones and only allows for collection and viewing of data, but we hope to expand features for our clients.
Specifically, we would like to automatically rate swings, using machine learning algorithms that analyze them.
Before implementing cloud technology, on-premise infrastructure would only age after installation. However, since cloud infrastructure can be constantly updated, we feel that is a big merit point in terms of security. In order to take advantage of this security merit, we are considering transferring the systems we have built over the years from local on-premise to the cloud.
In addition, since Layer 7 products are available, we hope to make use of programs such as Cloud Armor and Cloud Load balancing to increase the productivity of tasks that we have been doing manually.