vSAN 8 using the Express Storage Architecture (ESA) brings a new revolutionary way of exploiting modern hardware devices. This great alternative approach to achieving new levels of performance and efficiency requires new availability and serviceability improvements. This article will take a closer look at how ESA in vSAN 8 makes deploying and operating vSAN easier, and more flexible.
vSAN took a holistic approach by implementing enhancements in more than one area of the new architecture to reach a synergy effect in terms of serviceability and availability. Let’s check what are the features that have the greatest contribution to a better customer experience!
New resilient and efficient storage policies with no performance tradeoffs
Adaptive RAID-5 to accommodate cluster conditions
New levels of intelligence have been implemented into the storage policy engine for vSAN ESA. Now when a customer uses a RAID-5 erasure code policy rule, ESA will check for the host count of the hosts inside the cluster. Then it will decide which one of the two possible RAID-5 data placement schemes should be used. If the cluster has 6 or more hosts – a new 4+1 RAID-5 data placement scheme with parity components will be assigned. Respectively, if the cluster has less than 6 (3 to 5) hosts – a new 2+1 RAID-5 data scheme with parity components will be used. vSAN ESA will also have the capability to detect and determine if the number of hosts has changed. If so, it will change the RAID-5 policy rule accordingly based on the new host count. This new adaptive RAID-5 policy uses less capacity for storing data resiliently, bringing down costs.
Performance of RAID-5/6 equal to RAID-1
Performance of RAID-5/6 equal to RAID-1 is what customers can expect from vSAN 8 ESA. This can be accomplished by using FTT = 2, RAID-6 with the assistance of the new vSAN ESA’s log-structured filesystem and object format. Our recommendation is to use this storage policy whenever the cluster size is sufficient. This will provide the advantage of enhanced levels of resilience, and predictable levels of space efficiency, while not compromising on performance. Note that in vSAN ESA the objects will have a slightly different component layout than in the original storage architecture. If you want to learn more about the new components layout, check the “RAID-5/6 with the Performance of RAID-1 using the vSAN Express Storage Architecture” blog post.
Eliminating the disk group concept
Smaller failure domains.
vSAN 8 using ESA eliminates the disk groups concept which means that discrete failures of a storage device will only impact the given storage device, not the entire disk group. Servicing storage devices will also require less effort, since users do not need to understand the implications of a failed cache device, versus a failed capacity device, and how data services impacted the behavior of those failures. Additionally, this new type of configuration simplifies the disk claiming process while creating a host-based storage pool in vSAN 8 using ESA.
Furthermore, the removal of disk groups reduces the time needed for data resynchronization since it presents smaller failure domains, and it also leads to improvement in the availability SLAs.
VMware left no room for potential deployment and configurational complications when creating new vSAN ESA clusters by implementing new pre-checks into Cluster Quickstart to ensure customers are using approved hardware. It will check if the current hardware is compatible with the vLCM configuration; the minimum physical NIC link speed; the host’s physical memory and the type of disks that have been claimed. In addition, Skyline Health for vSAN will help ensure the cluster has been properly configured.
All these new features and enhancements are providing an overall easier to manage and more flexible environment. Customers should also feel safe that they will be properly guided on their new journey of deploying and operating the new vSAN 8 Express Storage Architecture.