We here at Geek+ don't take chances when it comes to Peak Season. We put our system through rigorous testing to ensure that no matter how slammed your warehouse gets with orders, our system will be able to handle it.
In anticipation for Peak Season 2022, we wanted to see just how tough our instantaneous order processing for our Goods-to-Person picking system was. We are excited to announce that our interface processing capacity reached 2.88 million orders per hour. That means, if there was a warehouse large enough to handle 2.88 million orders/hour, our IPS (interface platform system) could handle it.
What a number!
And if we built a platform to process a large amount of orders, we of course need to deploy large amounts of robots to fulfill those orders. After 8 years of tests preparing for peak cycles, not to mention years of algorithm and system-fine tuning, Geek+ standard Goods-to-Person picking solution software is able to schedule more than 2,000 robots in a single warehouse. This large scale reliability allows big brands to smoothly cope with peaks and has become a safe choice for large companies. The stability and efficiency of the picking system is something values Geek+ prides itself on, and the excellent performance of the picking system is inseparable from the continuous support of the system pressure test work.
What is a standardized stress test solution?
Every year in the early stages of Peak Season, Geek+ conducts a system stress test. We test the pressure resistance of the system in the face of large traffic, the high throughput of the system and the ability of the system to cope with failures, so as to ensure the stable operation of the system during Peak.
The stress test combined with fully automated stress testing tools, can
▸Achieve industry-leading pressure test efficiency and quality
▸Can better allocate basic service resources to the system
▸Accelerate outbound efficiency
▸Ensures outbound timeliness during the Peak period
▸Reliably and efficiently ensure the high stability of the system in the face of traffic peaks
The Pressure Test Process
The pressure test process of Geek+ combines single-step pressure test and full-link stress test methods.
Th process relies on the cloud environment to conduct a 1:1 simulation test of the system which simulates the situation of large traffic influx into the platform during Peak activity using about 5 times the estimated amount of normal business data. The system evaluates the traffic pressure that the platform can carry, and optimizes the system.
After 8 years of handling peak-like scenarios and more than 1,000 stress test projects, the standardized stress test solution formed by Geek+ empowers customers by:
▸Improving the concurrent processing capacity and outbound efficiency of the system
▸Fully ensuring the high stability and high throughput of the system in the face of large traffic
▸Quickly and accurately locates and solves system bottlenecks, greatly saving test time and costs, and easily meeting customers' emergency project stress testing needs
▸1:1 simulation flexibly adapts to scenarios with different business characteristics of customers
▸Relying on the real-time accurate parameter adjustment of the cloud environment
▸Flexibly responding to changes in order volume and improves throughput capacity
Single-step stress test
High concurrent processing capability to improve node efficiency
The single-step stress test is an initial performance test, that is, the efficiency and capacity test of the core link on each node, to ensure the stable and efficient operation of various functions including order preprocessing, wave, order on rebin wall and picking.
The system initiates high-concurrency calls to a single node in the test environment, analyzing its performance to evaluate the production efficiency of the system, finding the pressure-bearing bottlenecks of each link and optimizing them in time, eliminating foreseeable faults.
While stabilizing the system function, a high level of interface capability is also crucial. The interface of the picking system supports real-time feedback of picking information, which allows the upstream system to adjust orders in time according to the picking situation and perform operations such as cancellation or expedited picking. This year, the stress test results of the interface capacity exceeded 2.88 million /hour, the system function efficiency and capacity were higher and the system stability was stronger.
Full-link stress test
Optimize system resource allocation to cope with order peaks
Single-step stress testing is mostly independent, even if the indicators of each node are normal, the system link may not meet the standard after it is connected in series. Therefore, the stress test team will also conduct full-link stress testing of the system, aiming to find the overall performance problems of the system and continuously optimize it to ensure overall service performance results. The full-link stress test is also important to verify the ability of the system to resist the flood of orders. Compared with traditional offline stress testing technology, full-link stress testing has more accurate capacity prediction and improved performance, which can make up for shortcomings. The test can achieve greater business throughput, and ensure customer order processing capacity and outbound efficiency under the impact of large order traffic.
In addition, the full-link intelligent monitoring platform used by Geek+ can monitor business data to accurately analyze system performance, accurately adjust the system resource allocation, optimize the allocation of system resources, avoid expansion behavior, and ensure the efficiency and stability of the system.
1:1 simulation environment
High-fidelity stress test quickly adapts to customer production environment
Geek+ has undergone 8 years of peak order tests, using real business scenarios and online environments to obtain performance results that are more representative of actual business operations.
Using actual data from businesses, Geek+ continuously and effectively upgrades the test architecture, provides a test framework that quickly matches the customer's production environment, improves the quality and efficiency of testing, and effectively ensures system stability and high throughput. The smooth operation of the peak activities for many years has confirmed the accuracy of the 1:1 simulation results of Geek+.
Flexible and accurate parameter adjustment, agility guarantees stability
As one of the first companies in the industry to deploy a cloud environment, Geek+ has a long track record in cloud environment stress testing. The stress test team's optimization technology can be quickly adapted to meet customer needs. According to the estimated peak order volume provided by customers, the technology can accurately debug and configure various system parameters in real time.
The system can quickly deploy cloud resources according to stress test indicators accurately configured by the team, ensuring maximum throughput with the least resources.
We here at Geek+ believe that reliability in our products is most important. The more system stress tests we do, the less stressed our clients will be.