Technical Blogs
Network & Storage Quality of Service (QoS)

🗂️ Xloud XAVS – Network & Storage Quality of Service (QoS)


1. 🔎 Introduction

In high-density cloud environments, where multiple workloads coexist on shared infrastructure, performance isolation becomes essential.

Xloud XAVS delivers advanced Quality of Service (QoS) capabilities at both the network and storage levels—empowering enterprises to define, enforce, and monitor predictable performance tiers across their virtualized infrastructure.

In this guide, you will learn how XAVS enables:

  • Fair bandwidth allocation for tenant networks
  • IOPS control for critical and non-critical storage
  • DSCP-based traffic prioritization
  • Real-time monitoring of applied QoS rules via Grafana

2. 🧱 Architecture Overview

🔹 QoS Components in XAVS

ComponentRole
Neutron QoSDefines and enforces bandwidth rules, DSCP marking on ports
Cinder QoSApplies IOPS or throughput caps/guarantees to volumes
Cinder BackendsHosts multiple storage types (SSD/HDD) for different performance tiers
PrometheusCollects metrics for network and storage usage
GrafanaVisualizes bandwidth, latency, IOPS, and QoS effectiveness

🧩 These components enable end-to-end QoS enforcement across virtualized compute, network, and storage layers.


3. ✨ Key Features

  • 📶 Network-Level Bandwidth Control

    • Apply egress/ingress shaping per port
    • Use DSCP tags to prioritize traffic
  • 💾 Storage IOPS Management

    • Set read/write limits for volumes
    • Associate volume types with specific backends
  • 🏷️ Performance Tiers

    • Create Gold/Silver/Bronze tiers mapped to storage or network capabilities
  • 📊 Granular Monitoring

    • Visualize QoS performance metrics across tenants and tiers
  • ⚙️ Policy-Driven Enforcement

    • Define once, apply dynamically to VMs and volumes

4. 🧰 Use Cases

Use CaseDescription
Critical DatabasesAttach high-IOPS SSD volumes and guarantee bandwidth
Dev/Test EnvironmentsAssign mid-tier performance with moderate QoS limits
Archival StorageUse low-cost, throughput-capped HDD volumes
Multi-Tenant Bandwidth CapsPrevent bandwidth hogging by applying limits to noisy tenants

5. 🔗 Example Templates & Commands

🔹 A. Enabling Network QoS

✅ Step 1: Enable Neutron QoS

# /etc/kolla/globals.yml
neutron_plugin_agent: openvswitch
enable_neutron_qos: "yes"

✅ Step 2: Reconfigure Network Services

kolla-ansible reconfigure -t neutron

✅ Step 3: Create QoS Policy and Rule

openstack network qos policy create qos-limit-gold
 
openstack network qos rule create \
  --type bandwidth-limit \
  --max-kbps 100000 \
  --max-burst-kbps 50000 \
  qos-limit-gold
 
openstack port set --qos-policy qos-limit-gold <PORT_ID>

🔹 B. Storage QoS via Volume Types

🧱 Step 1: Define Cinder Backends

enable_cinder: "yes"
enable_cinder_backend_lvm: "yes"
enable_cinder_backend_ceph: "yes"
 
cinder_volume_backends:
  - name: gold_ssd
    volume_backend_name: GOLD_SSD
    volume_driver: cinder.volume.drivers.lvm.LVMVolumeDriver
  - name: bronze_hdd
    volume_backend_name: BRONZE_HDD
    volume_driver: cinder.volume.drivers.lvm.LVMVolumeDriver

🔄 Apply Configuration

kolla-ansible reconfigure -t cinder

🎛️ Step 2: Create Volume Types & Associate QoS

openstack volume type create gold
openstack volume type set --property volume_backend_name=GOLD_SSD gold
 
openstack volume qos create gold-qos \
  --consumer front-end \
  --property read_iops_sec=5000 \
  --property write_iops_sec=3000
 
openstack volume qos associate gold-qos gold

6. 🤖 Automation and Operational Fit

  • Ansible Automation

    • Use post-deploy playbooks to tag ports/volumes with QoS policies
    • Automate storage tier creation as part of provisioning workflows
  • Terraform Support

    • Integrate volume type and port creation via infrastructure-as-code modules
  • Grafana Dashboards

    • Pre-built panels for visualizing bandwidth, IOPS, latency per tier
  • Multi-Tenant Quota Templates

    • Apply default QoS templates based on customer class or billing tier

7. ✅ Summary & Positioning

XAVS QoS delivers fine-grained control over resource performance, enabling cloud operators to maintain service-level consistency even under contention.

By combining network shaping, storage tiering, and real-time monitoring, XAVS ensures:

  • Predictable performance for critical workloads
  • Fair usage enforcement in shared infrastructure
  • Scalable management of tenant expectations via policy

🟢 Customer Experience: Define your performance guarantees once and let XAVS enforce them at scale—across both network and storage layers.


© Copyright@xloud.tech