SBIR Application in Progress | Recruiting Early POC Partners

High-Concurrency Edge Governance Virtual Waiting Room

Designed for ticketing platforms, e-commerce flash sales, large-scale registrations, and limited-quantity events. Before traffic reaches the merchant’s website, the system handles queuing, filtering, traffic distribution, and controlled admission to reduce the risk of website outages, database locks, and unfair access caused by suspected scalpers and automated bots.

Are You Facing These Challenges?

The risk of a high-traffic event is not limited to server specifications. When a large number of requests simultaneously enter transaction workflows, membership systems, and databases, the entire service may become unstable.

Sudden Traffic Surges Overwhelm the Website

When concert tickets go on sale, major e-commerce promotions begin, or government registration systems open, traditional application-layer queuing may not react quickly enough to protect the primary website.

Suspected Scalpers and Bots Take Over Capacity

Malicious scripts, high-frequency requests, and automated behavior can occupy limited availability while increasing pressure on back-end systems.

Peak-Traffic Costs Are Difficult to Control

Businesses often maintain oversized server and database capacity throughout the year for only a few peak events, leaving resources underutilized during normal periods.

Understand How It Works in One Minute

Before users enter an event page or product page, their traffic first passes through the Cosmogrid Virtual Waiting Room edge governance layer. After validation, users are admitted in controlled batches according to the merchant’s configured capacity and admission rate.

01 | Large Volumes of Traffic Arrive

Regular customers, repeated refreshes, suspected scalpers, bot scripts, and event traffic arrive simultaneously.

02 | AI-Based Bot Behavior Filtering

Before traffic reaches the merchant’s website, it enters the edge validation layer. The system evaluates request frequency, behavioral characteristics, and risk rules to identify abnormal traffic.

03 | Waiting Room and Admission Rate Control

Validated users enter a fair queuing mechanism. The system releases users in controlled batches based on the merchant’s configured capacity and permitted admissions per second.

04 | Users Enter the Merchant Website

After traffic smoothing and validation, users proceed to purchasing, registration, checkout, or information retrieval workflows.

Core Technology and Product Capabilities

This solution does more than simply restrict traffic. It moves traffic governance outside the website, absorbs peak demand first, and then admits users according to predefined rules.

Edge Governance

Edge-Layer Traffic Governance

Move validation logic to the Edge, Nginx, or server entry layer so that requests do not directly reach the merchant’s primary website.

  • Pre-entry token validation
  • Waiting room and admission rate control
  • Reduced instantaneous load on the primary website
Fair Queue

Fair Queuing and Controlled Admission

Admit users into the production service in controlled batches based on event rules, back-end capacity, and the configured release rate.

  • Queue order management
  • Configurable admissions per second
  • Prevents large numbers of users from entering transaction workflows simultaneously
AI Bot Defense

AI-Based Bot Behavior Detection

Evaluate request frequency, behavioral patterns, and abnormal characteristics to reduce the impact of malicious scripts.

  • Suspected bot risk scoring
  • Abnormally high-frequency request detection
  • Can be combined with existing WAF and verification mechanisms
POC Report

Test Records and Validation Reports

During the POC stage, we can compile test records, load-testing data, waiting-room performance results, and recommendations for production deployment.

  • Event traffic records
  • Waiting and admission statistics
  • Production implementation recommendations

Reduce the Hidden Costs of Peak Traffic

The virtual waiting room still generates edge-computing and cloud-service costs. However, it can help businesses avoid maintaining oversized primary website infrastructure throughout the year for short-lived traffic peaks.

Reduce the Need for Permanently Oversized Servers

Avoid maintaining overprovisioned application servers and databases throughout the year for only a small number of major events.

Reduce Sudden Database Pressure

Controlled admission reduces the risk of deadlocks and timeouts caused by large numbers of requests simultaneously entering search, checkout, or registration workflows.

Convert Peak Costs into Event-Based Costs

Pay a lower maintenance fee during normal periods and purchase traffic packages when events occur, allowing costs to better reflect actual usage.

Two Deployment Models: Fast Cloud Integration and Server-Side Validation

The same waiting-room, token validation, bot risk assessment, and admission-rate control logic can be implemented through different deployment models according to the client’s website architecture, DNS management, security requirements, and operational capabilities.

Deployment Model Integration Method Best For Key Characteristics
CDN / Edge Fast Deployment Use DNS configuration to direct traffic from an event page or designated domain to the CDN or Edge layer first. The waiting room performs validation, queuing, and traffic distribution before forwarding users to the merchant’s existing website. E-commerce, ticketing, and event websites that need fast implementation, lower operational overhead, and support for clearly scheduled campaigns. Suitable for POC and SaaS deployment models. Major changes to the existing website are usually unnecessary. Testing can begin with a specific event page, subdomain, or designated entry point.
Nginx / OpenResty Server-Side Validation Configure Nginx with Lua or OpenResty validation logic at the client’s existing server entry layer. The server performs token validation, waiting-room status checks, and admission control. Businesses with self-managed servers, private clouds, internal systems, specialized security requirements, or a need to retain complete control over their traffic entry layer. Suitable for private or partially private deployment. It can integrate with existing Nginx configurations, reverse proxies, and internal validation workflows, but generally requires more technical coordination and testing.
In general, the CDN / Edge model is recommended when fast validation and lower operational overhead are the priorities. If the client has strict security, internal network, or private-deployment requirements, the Nginx / OpenResty server-side validation model can be assessed.

Suitable Use Cases

Any service involving concentrated short-term traffic, event rules, fairness requirements, or system stability can be evaluated for implementation.

Ticketing and Events

Concerts, performances, sporting events, event registration, and lottery-based registration.

E-Commerce Flash Sales

Limited products, member days, flash promotions, new product launches, and lucky-bag campaigns.

Government and Large-Scale Registration

Online applications, subsidy registration, quota allocation, information lookup systems, and peak-traffic entry points.

Founding Partner Collaboration Program

This service is currently in the POC validation and early-partnership stage. A limited number of partners are being invited to participate in priority testing. The collaboration focuses on validating real traffic scenarios, deployment processes, pricing models, and overall product feasibility.

Period Pricing Details
Months 1–3 Free POC validation and test-environment implementation. Includes assistance reviewing the website architecture, event scenarios, waiting-room workflow, and basic test data.
Months 4–6 20% of Standard Plan Pricing Extended testing period. Only a reduced testing service fee and required cloud costs are charged to support continued validation of event traffic and rule adjustments.
First Production Year 60% of Standard Pricing Partners that proceed to a production agreement after completing the POC receive a discounted first-year collaboration rate.
Year 2 and Beyond 80% of Standard Pricing or a Fixed Project Rate Long-term partners may retain preferential renewal pricing. A long-term fixed discount may also be negotiated when the partner provides case-study authorization and comprehensive product feedback.

POC-Stage Pricing and Usage Model

The service remains in the early validation stage. Pricing is assessed according to website scale, deployment model, event traffic, peak concurrency, and custom integration requirements. The recommended model consists of three layers: a baseline maintenance fee, event traffic packages, and project-based assessment for large-scale events.

The following prices are reference ranges for the POC stage and do not represent final public pricing. Actual collaboration terms will be adjusted according to the implementation scope and event risk.

1. Baseline Maintenance Fee

The baseline maintenance fee covers the retention of waiting-room settings, domain or entry-point configuration, basic monitoring, event-rule storage, and pre-event technical support. Basic protection may remain active during normal traffic periods, but this allowance is not intended for major event peaks.

Maintenance Tier POC Reference Price Best For Normal Usage Range
Basic Approximately NT$3,000–5,000 / month Small online stores, event pages, and branded business websites. Recommended monthly waiting-room requests: approximately 100,000–500,000. Peak concurrency: approximately 300–500 users.
Growth Approximately NT$6,000–10,000 / month Medium-sized online stores, recurring event websites, and member-day promotional websites. Recommended monthly waiting-room requests: approximately 500,000–3,000,000. Peak concurrency: approximately 500–2,000 users.
Enterprise Project Assessment Ticketing, large-scale registration, government, financial, and private-deployment projects. Usage, concurrency, SLA requirements, and deployment architecture require separate assessment.

2. Event Traffic Packages

An event package refers to the cumulative number of users or requests entering the waiting room during an event, rather than the number of users online at the same moment. For example, if 500,000 users enter the waiting room during a one-hour event, the event would approach the upper range of a small event package.

Event Package POC Reference Price Cumulative Event Traffic Suitable Scenarios
S — Small Event Package Approximately NT$5,000–15,000 / event Approximately 50,000–500,000 cumulative waiting-room requests. Small promotions, brand campaigns, one-time registrations, and initial POC testing.
M — Medium Event Package Approximately NT$20,000–50,000 / event Approximately 500,000–3,000,000 cumulative waiting-room requests. Member days, limited-product releases, popular event registration, and medium-sized ticketing events.
L — Large Event Package Approximately NT$60,000–150,000 / event Approximately 3,000,000–10,000,000 cumulative waiting-room requests. Large ticketing events, high-demand product releases, government registration, and high-risk traffic events.
XL — Extra-Large Event Custom Project Pricing More than 10,000,000 requests or clearly defined high peak-concurrency requirements. Major concerts, nationwide registration systems, specialized security requirements, and high-availability projects.

3. Advance Registration Discounts

Large events require advance configuration of the waiting room, edge rules, monitoring, traffic-release strategies, and testing procedures. Registration and testing should therefore be completed before the event to avoid insufficient scaling or incomplete configuration caused by last-minute activation.

Registration Time Pricing Details
At Least 14 Days Before the Event Approximately 80% of the Event Package Price Suitable for medium and large events. Allows time for technical confirmation, load testing, and rule adjustments.
At Least 7 Days Before the Event Approximately 80%–90% of the Event Package Price Suitable for small and medium events. Allows sufficient time for basic configuration and testing.
Within 48 Hours of the Event Standard Pricing or Rush Assessment Acceptance depends on engineering effort, traffic risk, and the feasibility of cloud scaling. Service availability cannot be guaranteed.
The figures 500,000, 3,000,000, and 10,000,000 primarily refer to cumulative requests or users entering the waiting room during an event. Peak concurrency refers to the number of users waiting or online at the same time and affects deployment specifications, monitoring, and required safety capacity. Events with significant expected peak concurrency require additional assessment.

POC Collaboration Process

Step 1

Requirements Review

Review the website architecture, event type, expected traffic, security limitations, and implementation scope.

Step 2

Deployment Model Selection

Select AWS Lambda@Edge or Nginx with Lua / OpenResty according to the project requirements.

Step 3

Test Environment Implementation

Configure the waiting room, token validation, admission control, bot rules, and basic monitoring.

Step 4

Simulated Load Test or Small Event

Use load testing or an actual small-scale event to observe waiting-room performance, back-end pressure, and user experience.

Step 5

Data Review and Adjustments

Adjust admission rates, waiting-room rules, and abnormal traffic assessment methods according to the collected data.

Step 6

Production Collaboration Assessment

After completing the POC, evaluate production plans, event packages, annual agreements, or private deployment.

Technical Team Experience

CosmoGrid Technology focuses on high-concurrency system architecture and MarTech traffic governance. The project lead has extensive experience developing and operating large-scale distributed systems and has worked on high-request-volume, high-QPS, and traffic-platform restructuring projects. This project aims to productize that practical experience into a virtual waiting-room service for ticketing, e-commerce, and large-scale registration scenarios.

Frequently Asked Questions

Is This Already a Commercial Production Product?

The service is currently positioned as a POC validation and early-partnership project. Closed testing and small-event validation can be conducted first, followed by discussion of a production commercial plan based on the results.

Does the Existing Website Code Need to Be Modified?

Not necessarily. Some scenarios can be handled at the edge layer, reducing the need to modify the primary website. However, limited integration may still be required for membership, checkout, ticketing, or registration rules.

Can It Completely Stop Scalpers or Bots?

The objective is to reduce the impact of suspected bots, abnormally high-frequency requests, and automated scripts. The service does not guarantee that every attack will be blocked.

Does the Website No Longer Need Additional Server Capacity?

No. The merchant must still maintain reasonable primary website capacity. The value of the virtual waiting room is to make traffic peaks more controllable and reduce the need to permanently overprovision infrastructure for short-lived events.

Preparing for a High-Traffic Event, Limited Sale, or Large-Scale Registration System?

Apply for early POC testing. We will first review your website architecture, event scenario, and estimated traffic, then recommend an appropriate deployment model, testing scope, and collaboration terms.

CosmoGrid Technology | Cosmogrid
Contact: Sky Chuang | Email: sky@cosmogrid.com.tw