ZebClient
Zebware.comCustomer Portal
  • ZebClient Documentation
  • Product Overview
    • Introduction to ZebClient
    • Functional Overview
    • Product Components
    • Data Flows
    • Data Striping
  • USE CASES
    • ZebClient Analytics
      • Architecture
        • Data Pipeline
        • Data Storage
        • Consumption
      • Deploy a ZebClient Advanced Cluster
        • Step 0 - Setup
      • FAQ
  • Planning & Getting Started
    • How to Choose your Deployment Mode
    • Deployment Modes and Tuning
    • Performance
    • License
    • Order Your ZebClient License
    • Pricing
  • Installation
    • Types of Installations
    • Guided Installation
      • ZebClient with Azure Blob Storage
        • Defining Backend with Azure Blob Storage
        • Mounting ZebClient with Azure Blob Storage
      • ZebClient with AWS S3
        • Defining Backend with AWS S3
        • Mounting ZebClient with AWS S3
      • Mount Additional Agent Node
    • Kubernetes
      • Azure Quickstart Guide
      • ZebClient CSI
      • ZebClient Helm
      • ZebClient Terraform
    • Virtual Machines
      • Azure Installation Guide
        • Installing Using Terraform
        • Uninstalling Using Terraform
    • Checking Installation
    • Running First Test
  • Management HOW-TOS
    • Add a New Agent VM into an Existing Cluster
    • Retrieve Cluster Log Files
    • CloudFormation Deployments
      • Understanding our CloudFormation Template
      • Uninstalling Using CloudFormation
    • Command-line Interface
  • Operations & Monitoring
    • Importing Your Data
    • Inlets
      • Data from External S3 Bucket
    • System Recovery Guide
      • Restore KeyDB Backup from S3
    • Port a Deploy
    • Add Resources to a Cluster
      • Add Application Node to Existing Machine (Manually)
      • Add Application Node to Existing Machine (via zc-cli)
      • Add New Application Node
      • Add Jumpbox
    • Verifying License Validity
    • Monitoring Your ZebClient Cluster with Netdata
Powered by GitBook
On this page
  • Reference of Test Results
  • Batched Reads and Writes
  • Small File Performance
  • Cluster Configuration Used for the Example
  1. Planning & Getting Started

How to Choose your Deployment Mode

PreviousFAQNextDeployment Modes and Tuning

Last updated 1 year ago

You can use this guide to decide what type of deployment will be best for your performance needs.

Using ZebClient on premises requires a local acceleration layer to compensate for the delays caused by the network uplink. Using ZebClient in the cloud allows you to take advantage of the cloud data center network infrastructure. When the ZebClient components and cloud storage are placed in the same location, you receive ideal conditions for your cluster.

Reference of Test Results

Below, you can explore our test results from sample workloads and configurations.

Batched Reads and Writes

Batched workloads require maximum performance in both read and write operations. The test first reads and then writes a substantial volume of data, which is a workload typical for ETL processes or batch training scenarios.

Small File Performance

This test is a race between clusters who can write and read 1000 4k files the fastest. Converged cluster deals with this matter in the most effective way, keeping up with much more expensive Tiered in write operations and surpassing it in read operations due to closely-knit architecture.

Cluster Configuration Used for the Example

  • Cloud provider: Azure

  • Agents: 3x Standard_L32s_v3

  • Database: Standard_D4s_v4

  • Blob Object Storage

  • (Tiered mode only) Servers 4x Standard_L8s_v3

Throughput
IOPS
IOPS