Trax Technology Solutions Pte Ltd., is a leading provider of in-store execution monitoring solutions for the fast moving consumer goods industry (FMCG). Utilizing proprietary image recognition technologies, Trax provides its partners with real-time in-store data more accurately, efficiently and cost effectively than ever before.
Sales personnel armed with mobile devices photograph retail shelves and upload pictures to Trax’ servers in the Amazon AWS cloud. There, SKUs are identified and data analyzed.
Within minutes, sales personnel receive live reports for action to be taken immediately while in store. Data is also transferred to HQ for further analysis, supervision and control. Trax systems provide immediate feedback on out-of-stocks, distribution, assortment, share of shelf, competition analysis, POS materials, exhibitions and more.
Data leakage is a frequently encountered issue in the cloud. With the, now, common practice of many companies storing their entire databases in cloud repositories, exposure to competitors is a challenge faced by all companies operating in the cloud, Trax included.
Visibility and Access Management
AllCloud employs a number of methods to ensure Trax with data security. Utilizing a Nagios based central monitoring system, which provides a visual, real time map of the online operation, each individually secured environment can be seen in detail, minimizing leakage risks. In order to manage Trax’ large number of separate environments, the Trax cloud security system grants every user, or machine, full access, so as to maintain the private network management for monitoring, access and central logging.
Disaster Recovery and Backup
AllCloud’s in-house backup tools continuously track and operate in all regions, performing AWS Relational Database Service (RDS) dumps into Amazon Simple Storage Service (S3). The tracking mechanism then automatically adds new RDS instances to the backup lists, generating Amazon Machine Images (AMIs) on a regular basis, in efforts to avoid data loss and secure Trax’ cloud environment for mission critical deployments.
Global Distribution Presence
Many companies that have accomplished a significant presence around the globe are faced with the interrelated challenges of providing high performance services in multiple regions and optimal system operation, concurrently.
Time to Market is of the Essence
Providing customers with real-time analyses of shelf execution is an ongoing challenge for Trax, as is real-time feedback for any real-time response service. Fortunately, this is where the cloud’s agility comes into play.
With the aim of maintaining a positive global presence, automation is key. AllCloud’s engineers have automated everything from the supporting ground infrastructure up to application stack deployment and integration. In so doing, they possess the ability to rapidly modify cloud infrastructure in the least amount of time to market. Consequently, this continuous deployment has enabled Trax to automatically upgrade or remove AWS resources on demand.
In addition, AllCloud’s in-house `Puppeteer’ solution extends the Puppet configuration management tool to support this prompt and continuous deployment; completely integrating Puppet and the AWS Application Programming Interfaces (APIs) as a Configuration Management Database (CMDB). This approach enables the dynamic cloud inventory and the attached application stack components to be tracked, including all of their detailed configurations.
Trax’ internationally distributed system and solutions, quick provisioning and extensive use of AWS cloud resources, create a high level of sophistication in regards to the release and deployment of the application stack. Moreover, there is a crucial need to adhere to a strict SLA that excludes downtime, yet incorporates operations across the cloud, in addition to smart access management.
“As a purely cloud Managed Service Provider (MSP), AllCloud specializes in capacity management, shutting down idle environments based on their individual schedules, which has the potential to be extremely cost effective if done intelligently. Co-hosting applications’ various components is used to consolidate assorted sections of the application stacks into fewer computing resources.”
Lahav Savir, CTO at AllCloud
Achieving an efficient cloud environment with maximum flexibility and minimum costs is no easy feat. Key procedures must be carried out, such as automation and continuous deployment (CD), in order to stretch an environment to its limits and beyond.
Leveraging Spot Instances
Trax’ uploaded picture analysis demands a great deal of computing power using M2 XLarge EC2 instances. The challenge here lies in delivering quick results in near real time, when thousands of pictures are received at once.
This process can be overcome by creating custom CloudWatch metrics that are updated at a high frequency which, by default, surmounts the low SQS built-in 10-minute delay. The system works to overcome latency whilst enabling cross-regional auto scaling, reaching the best zone and regional price. This is accomplished by utilizing AllCloud’s custom made monitoring system that tracks spot market prices, triggering an automatic rebound to the on-demand instance when a price is higher than the threshold.
Resizing, Reserving, Shutting down and Co-hosting
As an additional means to better ensure capacity management, AllCloud’s DevOps Central platform focuses on resizing machines on a recurring schedule, downsizing large instances when the demand is low. Furthermore, the system continually tracks and analyzes every instance’s weekly usage profile, using those findings to purchase a reserved EC2 instance which best suits each customer or application, by means of AWS Trusted Advisor.