Industries

Gaming Technology

We modernise and operate game backends at real scale: 100 million monthly players, 4 million concurrent, 300K auth requests per minute at peak. Erlang/OTP native expertise, AWS Advanced Tier Services Partner.

Talk to an engineer See shipped work →

Backends that survive a launch weekend

Game server infrastructure fails in a specific way: everything works in soak tests, then launch day multiplies traffic and the weakest dependency folds. Appliscale is a game backend development company with Erlang/OTP as a native competency and 11+ years of distributed systems work behind it.

Our engineers embed in existing game teams rather than bolting on a parallel org. We take services from design document to production deployment, monitoring and on-call, and we keep client identities confidential by default. For the full gametech practice, see gametech.appliscale.io.

100M

Monthly active players served

Concurrent players

300K+

Auth requests/min at peak

20M

Daily leaderboard queries

Messaging and social systems at 100 million players

For one client we modernised the chat and notification platforms used by multiple game teams. The work included upgrading Erlang service runtimes, migrating storage from Riak to Redis and Aurora, and moving deployments from raw EC2 to Kubernetes.

The result: systems that scale to 100 million monthly active users and up to 4 million concurrent players, with an estimated 50K USD per month cut from the AWS bill and fewer incidents pulling engineers off feature work. Read the details in the Erlang messaging services case study.

Player identity and authentication at scale

Identity is where game backends concentrate risk: one slow login service degrades every title on the platform. We run a cluster of identity work for major gaming platforms.

Authorization server. A phased migration from a debt-laden TypeScript service to Kotlin and Spring Boot, now handling 300K+ requests per minute at peak, hundreds of millions per day, with vulnerable dependencies removed along the way. See the gaming authorization server case study.
Login platform. Services for 15M+ daily active users, migrated to Spring Boot 3.2 and Java 21. A Redis caching layer cut S3 costs about 40 percent, load-test-driven downscaling cut operational costs about 30 percent, and monthly invoices dropped 40 percent. Details in the login application case study.
Unified MFA. An API-first MFA Factor API supporting mobile push approvals, TOTP and third-party authenticators, delivered by engineers embedded across backend and client SDK teams so it works consistently on PC, mobile and console. See the multi-factor authentication case study.

This is the kind of work engineering leaders bring us in for; if that is you, see how we work with CTOs.

Live ops infrastructure and real-time game data

Live ops runs on data paths that cannot lag. We built a game-agnostic leaderboard platform on Aurora MySQL, Redis and Kafka that serves 20 million daily queries at millisecond latency for 2.3 million daily users, with seasonal resets, tie-breaking, team boards and historical retention, integrated with two worldwide live multiplayer games.

For a US gaming company we broke a Groovy monolith into Go microservices over six months, delivering a transaction engine on EKS with ArgoCD-based GitOps that processes millions of requests per day. For Novomatic we designed a stateless Complex Event Processing engine that analyses 100K+ betting requests per minute and returns player incentives in near real time. Event pipelines like these are core data engineering work.

Erlang/OTPGoKotlinJavaRedisAuroraKafkaKubernetesTerraformAWS

Game backend cost optimization and reliability

Cost work in gaming is mostly architecture work. We remapped large Terraform stacks into Kubernetes by standing up a parallel environment, load testing it, and shifting traffic with Route53 weighted policies, so the migration caused zero service interruptions and unified runtime stacks cut cloud costs 15 percent.

We also standardise telemetry for gaming clients, including an ongoing engagement with a large global gaming company, because you cannot control cost or reliability you cannot see. That practice is described under telemetry and observability.

Talk to an engineer

Our Work

Proof, not promises.

All case studies →

Erlang Messaging Services

Cut AWS costs by $50K/month; scaled to 100M users, 4M concurrent

Reduced the client’s AWS costs by an estimated $50,000 per month while increasing the scalability of their systems to handle 100 million monthly active users and up to 4 million concurrent users

AWSErlangGoJava

Gaming Authorization Server

Handles over 300k of requests per minute peak, totalling hundreds of millions of requests daily

A service capable of handling around over 300 thousand requests per minute at peak time, totalling hundreds of millions of requests per day

Gaming Leaderboards

20 million daily queries with millisecond response times and 2.3 million daily users

A scalable, game-agnostic leaderboard handling 20 million daily queries with millisecond response times and 2.3 million daily users

Gaming performance and scalability

Processing millions of requests per day

Cut S3 costs by 40%, operational costs by 30%, invoices by 40% monthly

A 40% reduction in S3 storage costs, a 30% reduction in operational costs, reduced monthly invoices by 40%

Multi Factor Authentication

Unified MFA system through cross-team collaboration

Novomatic - Player incentives

Delivers near real-time analytics, processing 100K+ betting requests/min

Offers near real time analytics while processing hundreds of thousands of betting requests per minute

Streamlining gaming infrastructure

Cut cloud costs 15% with optimized infrastructure and unified runtimes

A 15% reduction in cloud costs through optimised infrastructure and unified runtime stacks

Frequently asked questions

The questions engineering leaders ask first.

Clear answers before a discovery call.

What concurrent player counts have you handled?

Our largest gaming engagement runs chat and notification platforms serving 100 million monthly active users with up to 4 million concurrent players. On the identity side we operate login services for a platform with over 15 million daily active users and an authorization service peaking above 300,000 requests per minute, which totals hundreds of millions of requests per day.

Do you work in Erlang and Elixir?

Yes. Erlang/OTP is a native competency, not an add-on. We maintain and modernise large Erlang messaging systems in production, including runtime upgrades and migrations from Riak to Redis and Aurora. Alongside BEAM languages we ship Go, Kotlin, Java and Scala services, so we can meet a polyglot game backend where it is.

Can you modernise a live game backend without downtime?

That is our standard mode of work. We migrated large Terraform-managed stacks from EC2 into Kubernetes by building a parallel environment, load testing it, then shifting traffic gradually with Route53 weighted policies so players never noticed. Auth and login migrations follow the same pattern: phased rollouts, contract tests, and careful canarying on services handling hundreds of millions of requests a day.

How do you cut game infrastructure costs?

By engineering, not by procurement alone. Modernising one client's messaging platforms saved an estimated 50,000 dollars per month on AWS. On a login platform we cut S3 costs about 40 percent with a Redis caching layer, cut operational costs about 30 percent by downscaling over-provisioned services after load testing, and reduced monthly invoices by 40 percent overall. An infrastructure unification project cut cloud costs a further 15 percent.

Do you provide on-call support?

Yes. We take on-call ownership for the services we build or adopt, including off-hours coverage that lets US-based client teams stay focused on roadmap work. For the leaderboard platform we run, hotfixes typically ship the day after an issue is identified.

Got something hard to ship?

Bidders, multiplayer infra, agentic platforms, or all three, tell us what you're building.

Start a conversation info@appliscale.io