Huawei Cloud to AWS in under 3 months, zero player impact.
PlayZap's global multiplayer platform, migrated end-to-end onto AWS with Novosoft — delivered ahead of schedule, 100% service availability, and a foundation for SEA, Europe and US expansion.
Customer background
PlayZap is an innovative technology company focused on casual game development. Built around the core idea that "games are social", PlayZap operates a portfolio of popular cross-platform multiplayer competitive games, with players across Southeast Asia, Europe and the United States.
To accelerate global expansion, the company decided to fully migrate its cloud infrastructure from Huawei Cloud to AWS. The move had to overcome cross-cloud service adaptation, a tight migration window, no in-house AWS experience, and heavy operational pressure on self-built services. With Novosoft as the delivery partner, PlayZap landed a smooth business migration and a modernised architecture — the foundation for the next phase of global growth.
Business challenges
PlayZap brought five interlocking problems to the table — each one capable of derailing the timeline on its own.
- Cross-cloud service adaptation. Core game services had been deployed on Huawei Cloud managed services for years. Subtle differences in functional coverage between Huawei Cloud and AWS meant full-process service adaptation, migration-feasibility studies and method validation were all required up front.
- Tight migration window, no spare hands. Leadership set a hard 3-month window with minimum impact on the live game business and pushed for early delivery. The customer's engineers were already fully committed to keeping the existing service running — there was no headcount to spare for migration work, so the entire process leaned on partner technical capability.
- High AWS learning curve. The team had no standardised AWS experience and was unfamiliar with AWS deployment and debugging conventions. With the migration window limited, there was no realistic path for the team to self-learn AWS deeply enough to ship game workloads on their own.
- High operating cost & instability of self-built services. Self-built MySQL and Redis on Huawei Cloud consumed significant ongoing maintenance time, drove up operating cost, and carried stability risk that threatened core scenes — login, matching, social. PlayZap wanted to move fully to AWS managed services to release that pressure.
- Technical groundwork for overseas expansion. The core business was in Hong Kong but the casual-game roadmap was global. AWS's global node coverage, technical maturity and partner ecosystem made it the right platform — and the migration needed to leave the operations and R&D team genuinely fluent in AWS, ready for SEA / EU / US rollout.
Novosoft's core solution
Novosoft ran the migration end-to-end and rebuilt the architecture around AWS managed services rather than re-platforming Huawei equivalents one for one. Every layer — edge, compute, data, observability — was redesigned to scale globally.
Global access & network layer
- User access. Amazon Route 53 for global DNS, combined with Amazon CloudFront CDN to accelerate static-resource delivery and cut latency for players in SEA, Europe and the US.
- Traffic distribution. Application Load Balancer (ALB) receives global player traffic and routes it to front-end / back-end services, keeping login and matching highly available.
- Network isolation. Public / Private subnets carved out of Amazon VPC. JumpServer and NAT Gateway sit in the Public subnet; core business and data services live in the Private subnet for security isolation.
Compute & application layer
- Elastic computing. Game front-end / back-end services run in EC2 Auto Scaling groups so traffic spikes from events and new game launches absorb automatically; compute-intensive game logic and pattern services run in Amazon EKS for rapid deploy and scale.
- Service layering. Front-end handles player interaction requests, back-end connects to the database and cache, and the EKS cluster carries compute-heavy game logic — keeping the system modular and scalable.
Data & caching layer
- Managed database. Amazon RDS in a multi-AZ 1-primary / 2-replica topology — primary handles writes, replicas absorb read load and disaster recovery. Automatic backups secure the data, fully replacing self-built MySQL.
- Distributed cache. Amazon ElastiCache for Redis with read / write separation. Write nodes accept real-time game data; read nodes support high-concurrency queries — eliminating the stability and ops pressure of self-built Redis.
- Data synchronisation. RDS and ElastiCache stay in sync, ensuring consistency and low-latency access to game state and player data.
Monitoring & operations layer
- Full-link monitoring. Amazon CloudWatch collects metrics and logs from game services, infrastructure and player behaviour, with alarm rules in place.
- Alarm notification. CloudWatch alarms trigger EventBridge, which fans out SMS / email through SNS + Lambda — fast incident response for the ops team.
- Audit & compliance. AWS CloudTrail records every API operation; S3 stores logs and Elasticsearch indexes them to meet ISO and GDPR audit requirements.
Data storage & analytics layer
- Logs & backups. Game logs and backups land in Amazon S3 and migrate automatically to cheaper storage tiers via lifecycle policies.
- Data analytics. Amazon Elasticsearch Service ingests S3 log data to power player-behaviour analysis and operational reporting — feeding the product iteration loop.
AWS services deployed
Revenue & performance results
- Ultra-fast migration, zero business impact. Full migration completed early — inside the 3-month window. The cutover from Huawei Cloud to AWS was zero-interruption and player-invisible. Login, matching, social and payment ran at 100% availability throughout; no player complaints, no migration-driven revenue loss.
- Cross-cloud adaptation, much stronger architecture. Every Huawei Cloud service was precisely adapted to its AWS counterpart. After migration the core business runs on AWS managed services — DB and cache availability at 99.99%, roughly 90% more stable than the previous Huawei Cloud self-built stack — fully closing out the stability risk and ensuring a smooth experience for global players.
- Operating costs down sharply, productivity released. Shifting self-built to AWS managed cut ops-labour cost by 60% (about $48K/yr saved), reduced core data-service operating cost by 35%, and lowered overall cloud-resource cost by 28% vs. Huawei Cloud (about $65,200/yr saved). Freed from infrastructure babysitting, the team's R&D and ops productivity climbed 40%, and the new-gameplay launch cycle shortened by 50%.
- Team fluent in AWS within 1 month. Customised training and on-site support handed the customer's engineers the keys to AWS — independently running deployment, ops and troubleshooting on the core products inside a month of cutover. Standardised AWS capability, in-house.
About Novosoft
Over a decade of IT excellence. For more than 10 years Novosoft has been at the forefront of the IT industry, turning complex challenges into competitive advantages for our customers. Our long-standing history isn't just about time — it's a deep reserve of expertise, resilience and proven success. We don't just follow industry trends; we help define them. Our extensive tenure ensures the solutions we deliver are cutting-edge, stable, secure and built to last.
Proven reliability
Consistently stable delivery, punctual completion and long-term trust.
Specialised expertise
Focused on core fields to provide deep insights and optimised solutions.
Have a similar move ahead?
From Huawei or GCP to AWS — or any cloud combination — Novosoft handles the migration end-to-end.
Talk to us