VMworld 2015: Virtual SAN – Technical Deep Dive & What’s New #STO5906

This session is/was true to its title and definitely dove deep into Virtual SAN (VSAN). Due to the extreme nature of details, requirements, parameters, etc, I decided to conclude the live notes about 40 minutes into the presentation. With much respect for court reporters and typists, finishing out the slide notes would have been of no more value that practicing my typing skills.

VSAN looks promising and maturing as a solution. While the concept of metro stretched clusters sounds very intriguing, I believe it is only practical in the right use cases. My own environment, for example, involves significant writing with extended operations, which would not be feasible to replicate live. Local performance would suffer greatly while database crunching generated large amounts of data requiring acknowledgement from the remote site before proceeding.

On the other hand, if your environment is web-scale or low-write intensity, then VSAN stretched clusters may offer great value to you. As always, it depends.

The closing consideration is sheer cost of a VSAN solution. The “HY-4” recommended starting point retails around $10-15K per node (read: $40-60K for the HY-4). That is hardware only, so vSphere and VSAN licensing costs pile on top of that.

The beta preview with dedupe and erasure coding for space efficiency may take VSAN to the next level and make even its premium cost more palatable. IMO: external storage is still the path until this possibility brings down the cost (assuming capacity, not compute, is the limitation).

Live Notes & Commentary

Agenda

  1. Introduction
  2. Customer Challenges
  3. SDDC & HCI
  4. Virtual SAN Overview
  5. What’s New in 6.1
  6. New Feature Details

Customer Challenges

  • Specialized Expensive HW: 120TB high-end hybrid SAN = $1.2M
  • Low Performance & Inflexible: 73% of enterprises lose productivity as a result
  • Complex Operations: typical 5 weeks to configure, test & deploy

SDDC & HCI

HCI: The Ideal Architecture for SDDC

VMware HCI: Single, Seamlessly Integrated Software Layer

  • Scale-out model on HC solution

Multiple Paths to HCI

  • Virtual SAN w/ vSphere + Virtual SAN Ready Nodes
    • Certified to run VSAN
  • VMware EVO:RAIL

Use Cases

  • VDI: flash
  • BCA: SQL, Exchange, etc
  • IT Operations
  • ROBO: 2-node config now available

Virtual SAN Overview

  • SDS optimized for VMs
  • HC architecture
  • Runs on standard x86 servers
  • Pools HDD/SSD into a shared datastore
  • Delivers enterprise-level scale & performance
  • Managed through per-VM storage policies
  • Deeply integrated with the VMware stack

Tiered Hybrid or All-Flash Architectures

  • Hybrid: uses SSD as read & write cache
    • 40K IOPS per host (*4K 70/30)
  • All-Flash: writes cached first; reads go directly to capacity tier
    • 90K IOPS per host w/ <1ms latency*

What’s New in 6.1

  • Enterprise Availability & Data Protection
    • Stretched cluster w/ RPO=0, metro-distance
    • 5-minute RPO vSphere Replication
    • Support for SMP-FT
    • Support for Oracle RAC & Microsoft MSCS
  • Advanced Management & Troubleshooting
    • Health check plug-in for hardware monitoring, compliance
    • vRealize Operations integration for capacity planning & RCA
    • Support for cloud-native apps (CNAs)
  • New Hardware Options
    • 2-node clusters for ROBO
    • New Ready Node platforms
    • New SSD HW options:
      • Intel NVMe
      • Diablo Ultra DIMM

Virtual SAN: Stretched Clusters

  • Overview
    • Site-level protection
    • Architecture is based on Fault Domains
    • Witness VM resides on a 3rd site
    • Automated failover in case of site failures
    • Scale to 15+15+1 nodes
    • Requires <5ms latency & >=10GbE bandwidth
  • Benefits
    • Disaster avoidance
    • Planned maintenance
    • Zero RPO

Virtual SAN 2-Node ROBO Solution

  • Overview
    • 2-Node solution
    • One witness VM per VSAN cluster
    • All sites could be managed centrally
    • ROBO bundle pricing per VM
  • Benefits
    • Cost effective
    • Space efficient

Virtual SAN Support for Business Critical Applications

  • Support for Exchange DAG
  • Support for SQL Server AlwaysOn Availability Group
  • Support for Oracle RAC

Virtual SAN Monitoring & Management Overview

  • Management tools
    • vRealize Operations Management Pack
    • Health Check Plugin
    • VSAN Observer
    • VSAN Assessment + Sizing & TCO calculator

IMG_4431

Virtual SAN Ready Node – New Branding

  • HY-2 Series
  • HY-4 Series
  • HY-6 Series
  • HY-8 Series
  • AF-6 Series
  • AF-8 Series

Virtual SAN: NVMe

  • Highly optimized controller interface
  • Improves performance for enterprise workloads
  • Provides increase performance over traditional SSDs

Virtual SAN: Ultra DIMM

  • Connects flash storage to memory channel
  • Enables <5 microsecond write latency

Feature Details

Stretched Clusters

  • Use Cases
    • Planned Maintenance
    • Disaster Avoidance
    • Automated Recovery
  • Architecture based on Fault Domains
    • Like any other VSAN cluster with 3 FDs
    • No VMs running on witness
    • Witness only holds meta-data
    • Witness VM resides on a 3rd site
    • Run the hosts at 50% of max VMs
  • Network Requirements
    • Max network latency of 5ms RTT
    • Dedicated 10Gbps for VSAN traffic recommended
    • For 5+5+1 config on server medium, network requirement is around 4Gbps (2Gbps each way)
    • Witness: L3 IP based, very low traffic, heartbeat once per second
  • Witness: reference vmware.com for documentation depending on scale
  • “Preferred” Fault Domain: used only when network partitioning
  • Read operations source out of local site
  • Write operations must be acknowledged from both sites before ack to application

VSAN Stretched Clusters + VR + SRM

  • Regional high availability plus distance replication w/ recovery

VMworld 2015 | Wednesday | Virtual SAN – Technical Deep Dive & What’s New (STO5906)

Vahid Fereydounkolahi, VMware | Rakesh Radhakrishnan, VMware

One Comment

Leave a Reply