At first, VisualCamp had two main pain points when entering China. First, they lacked the computing resources for running their deep learning models and frequently ran into development bottlenecks. Second, customers in China were dissatisfied with the speed and accessibility. Dealing with development and latency problems VisualCamp chose to migrate to Alibaba Cloud.

Featured Products

Elastic GPU ServiceApsaraDB RDS for MySQLAlibaba Cloud CDN
01

About VisualCamp


Human subconscious communication is often shown through various non-verbal means. Nonverbal communication, also called microexpressions, can often be more expressive than verbal communication. VisualCamp is a company whose proprietary AI can track eye movement for microexpressions, the gazing. It uses the latest in eye-tracking technology for VR, AR, and smartphones, and has a lot more uses than just telling if someone is lying. Used in a diverse number of applications such as content, retail, education, and advertising analysis, VisualCamp's eye-tracking algorithm can measure the gaze point of users by tracking pupil movement and use previous data to train their AI model further. VisualCamp believes eye-tracking technology has the potential to increase the quality of life for its users, especially for disabled people and patients who experience difficulty using digital devices. The ultimate goal of VisualCamp is to become a data platform company specialized in handling gaze data. To expand its business domain into more device types, operating a stable server for gaze data streamed from all different devices is essential. As China is a key market that has the most significant number of smartphone users, utilizing Alibaba Cloud for protecting Chinese users' gaze data and managing the analysis process is an essential part of its business.
02

The Challenge


At first, VisualCamp had two main pain points when entering China. First, they lacked the computing resources for running their deep learning models and frequently ran into development bottlenecks. Second, customers in China were dissatisfied with the speed and accessibility. Dealing with development and latency problems VisualCamp chose to migrate to Alibaba Cloud.
03

Deployment


The solutions Alibaba Cloud technicians came up with was to utilize several GPU cloud servers, which would increase development speed and be able to process deep learning algorithms. Besides, Alibaba Cloud Solution Architects set up instances in multiple locations in China and in Hong Kong to improve latency and better serve local customers.
04

Why Alibaba Cloud


Alibaba Cloud's Elastic GPU Service GN5 instance with NVIDIA Tesla P100 (world's first AI supercomputing data center GPU) satisfied VisualCamp's deep learning needs (as much as 3 to 5 million frame data per day). After setting up instances in Beijing, Shanghai, Hong Kong, Singapore, the development teams were able to deliver products to their customers faster. Using Server Load Balancer allowed VisualCamp to handle massive traffic with flexibility. Finally, VisualCamp's operational efficiency increased after switching to Alibaba Cloud. VisualCamp was able to add servers in urgent cases quickly, and an automatic snapshot every 24 hours meant that their data was backed-up and secure. By leveraging Alibaba Cloud's services, VisualCamp can now run and stop instances as needed, allowing for better allocation of resources, saving them both time and money.