Introduction

In 2024, over 2 billion people will participate in elections in 60+ countries, including India, Mexico, the United Kingdom, and the United States. To manage these elections, AWS customers are preparing their infrastructure. Cloud services play a vital role in the election process and have an increasing impact across the entire lifecycle of an election, from campaigning, voter registration, and vote counting to information broadcasting. To successfully prepare for these events, you must make sure that all the components of your infrastructure and applications are optimized to handle expected workloads. This article offers guidance and tips to help you prepare for the election season and protect related data with security and integrity. If you’re preparing for the upcoming elections or managing election-related infrastructure, including infrastructure engineering teams, customer engagement teams, cloud operations teams, and product managers, then this article is for you.

Election use cases

Elections are one of the most important administrative processes for democratic governments. However, elections often experience an array of sophisticated threats that are designed to undermine system integrity and public trust. It's critical to maintain transparency and trust in our election systems, especially when disinformation is rampant. Also, the risk of misinformation is higher currently because of the introduction of Generative AI. Based on past experience, the following three use cases are the top asks from customers during the election season:

• Live streaming: Manage streaming of election-related events for engagement, election results, and broadcasting. As the number of viewers increases, the underlying infrastructure must scale without any lag between live and streamed events, while also handling advertisement and analytics insertion.
• Manage voting systems: Assist with the security, integrity, and scalability of your voting systems to establish transparency and combat the increasing threat of cyber-attacks. The US Department of Homeland Security (DHS) declared the electoral system as "critical infrastructure". This places election infrastructure in the same category as the US power grid or financial sector.
• Track and publish election results: Design systems that support real-time data ingestion and perform analysis, and render up-to-the-minute election results for your audience. As the number of users increase, the infrastructure must scale automatically with no impact to performance in a cost-effective manner.

Solution overview

AWS Countdown Premium, an AWS Support offering, provides critical support across all phases of election from design to post-election event retrospectives. This service offers designated engineers that are selected from a team of AWS experts to do the following:

• Provide proactive guidance and troubleshooting with proven playbooks.
• Apply the best practices that are typically used for major events, such as Amazon Prime Day.
• Provide critical support at every step of your election planning and implementation, from design to post-launch retrospectives.

Designated engineers get involved from project inception to confirm continuity, provide access to subject matter experts, and leverage support tools for faster issue resolution. They participate in critical event calls for live streaming events, voter registration, and vote counting to provide rapid issue triage and resolution that’s equipped with application context. With AWS Countdown Premium, you can increase your infrastructure investment return by delivering high-impact events and achieving your organizational goals.

AWS Countdown Premium empowers election administrators and other stakeholders to focus on the core needs of the dynamic electorate in a secure, scalable, resilient, and cost-effective way. This service extends your team’s capabilities with designated cloud experts. These experts can guide you through the planning, preparation, and execution stages of your applications with best practices, AWS Well-Architected Reviews, Operational Readiness Reviews (ORRs), security reviews, and recommendations. That way, you can focus on your core needs instead of building and maintaining the underlying infrastructure to support your mission-critical efforts.

AWS Well-Architected Reviews

For the three previously mentioned election use cases, there’s usually a spike in user traffic during critical events windows. Vote publishing interfaces can receive over 50 times the peak traffic during US presidential elections in comparison to county elections. The increasing traffic and potential cyberthreat by adversaries increase the stress on your infrastructure, and can possibly result in slow response times and system failures. The AWS Well-Architected Framework review evaluates the architecture of your workloads against AWS best practices across six pillars: Operational Excellence, Security, Reliability, Performance Efficiency, Cost Optimization, and Sustainability. This in-depth review helps you build a secure, high-performance, resilient, and efficient application infrastructure. AWS Well-Architected Reviews identify risks, recommend remediation of those risks, and cover other areas that are critical to election-related workloads, such as confirming sufficient capacity, workload health, security, and optimized costs. Ideally, AWS Well-Architected Reviews must be part of an organization’s best practices and conducted at least 2 to 3 months in advance to allow time to implement recommendations.

Operational Readiness Review

Your voter engagement solutions might use services such as Alexa, Amazon Lex, and Amazon Connect to manage call centers or chatbots. Or, you might have an AI-driven social media dashboard that’s built with Amazon Comprehend. An ORR evaluates the readiness, availability, and risks that are associated with your election-related production workloads. The review then creates a series of action items with remediation to be addressed by your teams. During this process, an AWS expert reviews your operational mechanism, typically with your development, operations, and DevOps teams. The AWS expert evaluates and reviews the mechanisms for your operational goals, risk assessment, architecture assessment, monitoring, deployment, testing, team structure, and escalation procedures. You must perform this review at least 1 month before the event.

Security review

The security of voting systems is essential to the democratic process, with cybersecurity playing an increasingly vital role. AWS meets over 90 international compliance standards, certifications, frameworks, and authorizations, such as FedRAMP in the US and ISO 27001/27017/27018 internationally. AWS provides services that adhere to the high privacy bar and data protection standards that are required of data processors by the GDPR for data privacy in the European Union. AWS offers multiple services and resources to help you meet these same standards, certifications, and frameworks, and prepare you against malicious activities targeted against election systems. The Self-Service Security Assessment solution generates a point-in-time assessment of your security posture and offers actionable insights to triage risk points. You can also run this solution multiple times to assess ongoing progress of applied remediation. AWS Shield, a managed Distributed Denial of Service (DDoS) protection service, is used along with Amazon CloudFront and Amazon Route 53 to provide comprehensive availability protection against all known infrastructure (layer 3 and 4) attacks. AWS Countdown Premium engineers help you validate your security framework. They help you build a landing zone to make sure that your workloads are protected against DDoS attacks, are highly available, and are fault tolerant. You must perform the security review at least a month before the event.

Infrastructure scaling

The first step to scale traffic is to assess the expected traffic along with the current infrastructure capacity. In addition, it’s important to identify the event scope with a focus on resource constraints, third-party dependency, and disaster recovery options. This makes sure that all relevant resources can scale in correlation with increasing variable traffic. AWS supports scaling at both the individual and aggregate level with AWS Auto Scaling that adjusts capacity to maintain a steady performance at the lowest cost with variable traffic. Some services must be “pre-warmed” in case of steep traffic forecasts. You can reserve the required resources with self-service tools or by contacting AWS Support. Your AWS Countdown Premium engineers will assess your capacity requirements and help you reserve the infrastructure in the most cost-effective way. They will help you choose the best option among auto scaling, pre-warming, or a combination of both based on your systems and the expected traffic pattern. For example, the Federal Election Commission expects to review 1 billion campaign contributions on AWS in 2024 and will scale up the infrastructure to meet user demands quickly.

Monitoring

Monitoring is crucial to detect and rectify any issues that arise during an election event. An effective monitoring system tracks system performance, identifies potential bottlenecks, and promotes issue mitigation. You can track your metrics on vote publishing interfaces or live streaming on Amazon CloudWatch, or your chatbot performance in the Amazon Lex console. AWS Countdown Premium engineers help you identify critical metrics and develop dashboards, alarms, and mitigation plans to proactively mitigate or resolve issues. For example, your AI-driven social media dashboard can help you capture real-time customer feedback to help identify any issues with your systems.

Load testing

Load testing helps you understand election system behavior and benchmarks it against the expected peak traffic during elections. Exhaustive load testing across application, data, and infrastructure layers is important to build confidence in the systems architecture and operational processes to achieve your goals. Recently, a vote publishing interfaces experienced a higher latency because of increasing user traffic, resulting in poor experience. The issue was caused by an artificial limit on database connections in the application layer. Exhaustive load testing might have prevented this issue. AWS Countdown Premium engineers guide you on load testing and help you fine-tune alarm thresholds against your relevant resources. An escalation matrix and communication plan must be in place before the event and have owners assigned to each high-priority risk.

Post-event

After the election event is complete, you can ramp down your infrastructure resources. This is one of the many benefits of running your workloads on AWS. You can run a post-event retrospective to identify gaps in your processes and create a baseline for similar future events. During the retrospective, you can focus on incidents during the event window, runbook effectiveness, monitoring effectiveness, a communication plan, and the performance of dependent services. With AWS Countdown Premium, the designated engineer facilitates this process.

Conclusion

Successful election events rely on reliable, scalable, and secure systems, where all system layers work together to create the most effective and secure experience. The success of your event heavily depends on planning and preparation of reviews. Make sure that you spend sufficient amount of time to remediate and strengthen your infrastructure and workloads. AWS Countdown Premium can help you on every step of your preparation and execution.

To learn more about AWS Countdown Premium, see AWS Countdown Premium. To sign up for this service, open AWS Support Center. To learn more about how our plans and offerings that can help you get the most out of your AWS environment, see AWS Support.

---

About the author

Atul Anand

Atul is a Senior Product Manager at Amazon who has a unique blend of operations and product expertise. He brings a customer-centric approach to his work, leveraging a wide range of product experience across multiple domains to drive innovative solutions for customer problems.