# How it works — Fluxzero Source: https://fluxzero.io/how-it-works/ ## A cloud that takes care of itself. Enterprises expect critical systems to keep their promises under pressure. That standard matters long before a product reaches enterprise size. - **GB / min data throughput:** 100 + Or about 700k tweets per sec, nearly 5 times X.com's record. [*](https://fluxzero.io/how-it-works/#tweet-throughput-note) - **Requests / sec:** 1 M+ Completed requests, end-to-end across applications. - **Less code:** up to 99% Across legacy migrations with the same behavior. * Estimate based on a [2.3 kB average serialized tweet size](https://www.research-collection.ethz.ch/bitstreams/77b38ad3-f165-4bda-9fca-6a1d13913836/download) and [X.com’s published record of 143,199 tweets per second](https://blog.x.com/engineering/en_us/a/2013/new-tweets-per-second-record-and-how). 01 · Foundation by default ### Your product comes with the full Fluxzero foundation. Your applications and data run together in a dedicated environment called a cluster. Every cluster is isolated from every other cluster, and its applications stay off the public internet. From the first deployment, its foundation provides correctness, security, privacy, recovery, auditability, and controlled scale. These guarantees come with every cluster, so your agent can focus on the product. **External traffic** Your product cluster **Application A** 3 tasks **Application B** 2 tasks **Fluxzero foundation** Correctness Security Privacy Recovery Auditability Scale Every cluster isolates one product and includes the complete Fluxzero foundation from its first deployment. 02 · Correctness ### Every action leaves the product in a valid state. Your product defines what must always be true. Fluxzero makes sure those rules still hold when many users act at the same time. A change either succeeds completely or leaves nothing half-finished behind. - **Buyer A** **Last ticket** Available **Reserved** - **Buyer B** **Last ticket** Already reserved **Rejected** Both buyers attempt the same ticket. Only one reservation succeeds; the other is rejected. 03 · Reliability and recovery ### Failure does not mean lost work. Fluxzero records work before your application starts it. If an instance restarts, the work waits and continues when it is ready. One failing part does not stop the rest of the product. Fix the cause, then retry only the affected work. Failed work remains visible and recoverable without first building a separate recovery system. - 1 - 2 - 3 Failed - 4 - 5 04 · Privacy and data control ### Keep the history. Remove the personal data. Sensitive values stay separate from the durable history. Trusted code can use them without exposing them in logs, and they can be permanently erased without destroying the surrounding record. Fluxzero protects PII data (Personally identifiable information: data that can identify a person directly or indirectly, such as an email address.) and makes it erasable while you decide how it is used. Each recipient sees only the data the product allows. These controls make data minimization and erasure practical and support requirements under GDPR (The European Union’s General Data Protection Regulation, governing how organizations collect, use, and protect personal data.) in Europe and California’s CCPA (The California Consumer Privacy Act, amended by the CPRA, including rights to access, delete, and limit the use of personal information.). **Order 42** size M email Delete personal data **Order 42** size M redacted The protected email is erased while the rest of order 42 remains available. 05 · Secure by design ### Your applications stay off the public internet. Fluxzero receives public web traffic at its managed edge. Your applications pull incoming traffic and send outbound traffic over connections they open themselves, so none of them needs a public endpoint. Your applications still decide who someone is, what they may do, and whether a request is valid. Fluxzero gives every request a protected path to those product decisions. **Public internet** **Fluxzero foundation** Traffic in and out **Apps pull and send** Outbound connections Your cluster **Application A** No public endpoint **Application B** No public endpoint Fluxzero receives public traffic while applications stay private and pull and send traffic over connections they open themselves. 06 · Product insight ### Deep insight. Built in from day one. Fluxzero gives your agent all the evidence it needs to understand production. Every product starts with this depth of insight, without extra tools or setup. [See how Fluxzero gives you full insight →](https://fluxzero.io/product-insight/) 07 · Bounded traffic ### Traffic spikes don’t overwhelm your applications. Fluxzero limits how much work reaches each application at once. Excess demand waits outside instead of flooding it, so DDoS-like overload cannot cascade through the product. Excess work **Bounded capacity** Application 08 · Independent scale ### Each part of the product scales independently. Slow work cannot stall the rest of the product. Capacity grows exactly where it is needed. Start as one application, then split it into separate services—or bring it together again—without changing what the product does. Recorded events **Ticketing** **Payments** **Notifications** **One application** same product behavior **Independent services** 09 · Safe evolution ### New capabilities can reach into the product’s past. Your product can change without abandoning the data and decisions it has already recorded. Older information remains usable by the current application, even as the product evolves. Add a new capability later and run it across existing history. It can work retroactively instead of starting empty on the day it is introduced. **Audit trail** 18 days ago **Ticket reserved** 14 days ago **Payment received** 10 days ago **Reservation confirmed** Already running **Ticket reservations** **Payment processing** **Customer confirmations** **Cancellation handling** Added later **Customer rewards** Built from all previous purchases Ticket reservations, payment processing, customer confirmations, cancellation handling, and customer rewards can all use the complete product history. 10 · ZeroOps cloud ### Deploy. Pause. Restore. Deploy to as many clusters as your products require. Fluxzero operates every one from its first deployment, with point-in-time recovery backed by automated backups and a complete monitoring view available as soon as it starts. Create a new cluster or duplicate one from an earlier point in time. Each cluster can be paused without losing its identity or state, while access and secrets remain protected. **Production** Cluster **Acceptance** Cluster **Built into every cluster** - **Deploy** Start a new cluster - **Pause and resume** Preserve identity and state - **Point-in-time recovery** Automated backups; restore or duplicate - **Protected access** Secrets stay encrypted **Monitoring in every cluster** - **Audit trail** Every message and interaction - **Application logs** Search all application output - **Traces and issues** Follow causes and failures - **Insights and visualizations** Performance and trends Production and acceptance clusters share the same operated lifecycle, point-in-time recovery backed by automated backups, and complete monitoring view. 11 · A shared foundation ### The foundation keeps improving underneath every product. Every Fluxzero product runs on the same hardened foundation. When that foundation becomes stronger, every product benefits without taking on more infrastructure code. Your application remains the product itself. It does not become a permanent record of every technical decision made on the way to launch. **Your product** Behavior, rules, experience **Fluxzero foundation** Correctness Security Privacy Recovery Auditability Scale Continuously improved by the cloud Built into every product ### Less to manage. More to see. [Get started](https://fluxzero.io/#get-started) [Explore the docs](https://fluxzero.io/docs/)