> For the complete documentation index, see [llms.txt](https://docs.zetrix.com/zetrix-l2-documentation/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.zetrix.com/zetrix-l2-documentation/network-and-deployment/production-deployment.md).

# Production Deployment Recommendation

## Topology (Proposed 8-Server Separation)

For Mainnet, the team intends to **follow the same consolidated approach as Testnet initially**, then adopt the following component separation as scalability enhancements become necessary. This is the proposed target topology.

| Server      | Role(s)                                          | Baseline Sizing                | Public Domain                                 |
| ----------- | ------------------------------------------------ | ------------------------------ | --------------------------------------------- |
| Server 1    | Mainnet Node 1 + Sequencer 1                     | 2 core, 4 GB RAM (scale later) | — (behind NLB)                                |
| Server 2    | Mainnet Node 2 + Sequencer 2                     | 2 core, 4 GB RAM (scale later) | — (behind NLB)                                |
| Server 3    | Mainnet Node 3 + Sequencer 3                     | 2 core, 4 GB RAM (scale later) | — (behind NLB)                                |
| Server 4    | Mainnet Node 4                                   | 2 core, 4 GB RAM (scale later) | — (behind NLB)                                |
| Server 5    | Mainnet Node 5                                   | 2 core, 4 GB RAM (scale later) | — (behind NLB)                                |
| **AWS NLB** | Maps `z2-node.zetrix.com` to all Node server IPs | —                              | `z2-node.zetrix.com`                          |
| Server 6    | Main Proposer Validator + Reexecutor             | 2 core, 4 GB RAM (scale later) | `z2-validator.zetrix.com`                     |
| Server 7    | Explorer                                         | 2 core, 4 GB RAM (scale later) | `z2-explorer.zetrix.com`                      |
| Server 8    | Refueler + Relayer + Monitoring + Faucet         | 2 core, 4 GB RAM (scale later) | `z2-faucet.zetrix.com`, `z2-admin.zetrix.com` |

> \[!NOTE] All servers start at **2 vCPU / 4 GB RAM** and are **scaled vertically (or horizontally) later, when load requires it**. This deliberately mirrors the low-footprint Testnet deployment.

> \[!NOTE] In this 8-server proposal the 5-member **Committee (Validator + Reexecutor, Committee + Aggregator role)** co-locates with the Node servers (1–5); dedicated Committee servers can be added as the network hardens. Increasing the number of independent Committee and Watcher operators is the primary decentralization lever.

```mermaid
flowchart TB
    subgraph S1[Server 1]
        N1[Node 1]:::n
        SQ1[Sequencer 1]:::s
    end
    subgraph S2[Server 2]
        N2[Node 2]:::n
        SQ2[Sequencer 2]:::s
    end
    subgraph S3[Server 3]
        N3[Node 3]:::n
        SQ3[Sequencer 3]:::s
    end
    subgraph S4[Server 4]
        N4[Node 4]:::n
    end
    subgraph S5[Server 5]
        N5[Node 5]:::n
    end
    NLB[AWS NLB<br/>z2-node.zetrix.com]:::lb
    subgraph S6[Server 6 — z2-validator.zetrix.com]
        PR[Main Proposer Validator]:::p
        RE6[Reexecutor]:::r
    end
    subgraph S7[Server 7 — z2-explorer.zetrix.com]
        EXP[Explorer]:::x
    end
    subgraph S8[Server 8 — z2-faucet / z2-admin]
        REF[Refueler]:::x
        REL[Relayer]:::x
        MON[Monitoring]:::x
        FAU[Faucet]:::x
    end

    NLB --> N1 & N2 & N3 & N4 & N5
    S1 & S2 & S3 --> S6
    S6 --> L1[(Zetrix L1)]
    S8 -.->|top-up / relay| S1
    S8 -.->|top-up| S6

    classDef n fill:#dbeafe,stroke:#3b82f6;
    classDef s fill:#fef3c7,stroke:#f59e0b;
    classDef r fill:#f3e8ff,stroke:#a855f7;
    classDef p fill:#fee2e2,stroke:#ef4444;
    classDef x fill:#e5e7eb,stroke:#6b7280;
    classDef lb fill:#cffafe,stroke:#06b6d4;
```

## Infrastructure Sizing

All roles start at a uniform **2 vCPU / 4 GB RAM** baseline and scale later when required. Storage and network below are practical starting points; monitor and grow per [Monitoring](/zetrix-l2-documentation/architecture/system-components.md#monitoring).

| Server / Role                                 | vCPU (baseline) | Memory (baseline)  | Recommended Storage        | Storage Type | Network |
| --------------------------------------------- | --------------- | ------------------ | -------------------------- | ------------ | ------- |
| Node + Sequencer (S1–S3)                      | 2 (scale later) | 4 GB (scale later) | 250–500 GB (grows)         | SSD/NVMe     | 1 Gbps  |
| Node (S4–S5)                                  | 2 (scale later) | 4 GB (scale later) | 250–500 GB (grows)         | SSD/NVMe     | 1 Gbps  |
| Main Proposer + Reexecutor (S6)               | 2 (scale later) | 4 GB (scale later) | 100–250 GB                 | SSD          | 1 Gbps  |
| Explorer (S7)                                 | 2 (scale later) | 4 GB (scale later) | 250 GB–1 TB (index grows)  | SSD          | 1 Gbps  |
| Refueler + Relayer + Monitoring + Faucet (S8) | 2 (scale later) | 4 GB (scale later) | 250 GB (metrics retention) | SSD          | 1 Gbps  |

> \[!NOTE] Node and Explorer storage grow with chain history and indexing. Provision headroom and enable predictive disk alerts (see [Monitoring](/zetrix-l2-documentation/architecture/system-components.md#monitoring)). When vertical scaling of a Node server is required, prefer NVMe and increase RAM before CPU for state-DB workloads.

## Expected Network Bandwidth

* **Nodes:** highest — P2P gossip + RPC serving. Provision ≥1 Gbps, burst higher for public RPC.
* **Sequencer/Proposer:** moderate; bursts on batch/assertion submission to L1.
* **Monitoring/Explorer:** scales with scrape volume and query traffic.

## Scaling Recommendations

* **Read/RPC:** add stateless RPC nodes behind the NLB.
* **Reexecutors/Watchers:** scale horizontally; independent operators for safety.
* **Explorer/Monitoring:** scale API/UI and storage tiers independently.
* See [Scaling Strategy](/zetrix-l2-documentation/reliability-and-security/scaling-strategy.md#scaling-strategy) for detail.

## High Availability Recommendations

* 5-node QBFT quorum (tolerates 1 node loss).
* Active/standby Sequencer with fencing.
* Standby Proposer with leader election.
* Redundant Relayer/Monitoring/Explorer; replicated databases.
* Multi-AZ placement; automated failover.

## Deployment Diagram (HA, Multi-AZ)

```mermaid
flowchart TB
    subgraph AZ_A[Availability Zone A]
        N1[Node 1]
        N2[Node 2]
        SQA[Sequencer active]
        PRA[Proposer active]
    end
    subgraph AZ_B[Availability Zone B]
        N3[Node 3]
        N4[Node 4]
        SQB[Sequencer standby]
        PRB[Proposer standby]
    end
    subgraph AZ_C[Availability Zone C]
        N5[Node 5]
        SVC[Services: Refueler/Relayer/Monitoring/Faucet]
    end
    NLB[AWS Network Load Balancer] --> N1 & N2 & N3 & N4 & N5
    SQA -. failover .- SQB
    PRA -. failover .- PRB
    N1 & N3 & N5 --> L1[(Zetrix L1)]
```

***

[← Deployment Architecture (Current Testnet)](/zetrix-l2-documentation/network-and-deployment/deployment-architecture.md) · [Index](/zetrix-l2-documentation/readme.md) · [Networking →](/zetrix-l2-documentation/network-and-deployment/networking.md)