Today has seen the public release of Nutanix’s AOS 5.6 which contains fixes to code (as you’d expect from a software company) but also new additions of which the major ones are covered here.
Volume Group Load Balancing
There comes the point where performance demands need an injection from new software to boost the throughput, and in this release, it comes in the form of virtual disk load balancing. If we look at the Nutanix Controller VM one its roles is to service disk I/O which let’s just be clear it manages quite handsomely via the Stargate process, however, with new disk technologies such as NVMe and 3D XPoint there needs to be delegation and parallel processing for vast I/O requests – not forgetting of course that AHV Turbo has opened the door to really drive throughput for these new drive architectures.
(Detailed explanation on AHV Turbo on Josh Odgers website)
Enter VGLB. Using Acropolis Block Services (ABS released in AOS 4.7) and presenting storage to a virtual machine via a Volume Group means the virtual machine’s vDisk(s) will be split across multiple nodes in the cluster and accessed in parallel via multiple Stargate processes. Now the entire Nutanix Cluster can deliver the I/O for the VM using this feature.
2 Node Cluster support
Using the NX-1175s models, a 2 node Nutanix Cluster is now supported. As discussed in the post here this model and deployment use case is for Remote Branch Office environments and only available on the NX platform on day 1.
A witness role is provided by a single virtual machine (VM), this can be hosted anywhere and has no bandwidth/latency requirements. Additionally, this witness VM can be responsible for multiple sites.
Two-node clusters are only available in AHV and ESX environments and cannot be expanded from the initial 2 to 3, 4, etc… node clusters.
Flow Micro-segmentation (General Availability)
In AOS 5.5 micro-segmentation was released in “Tech Preview” mode and only supported for test and development use cases. The Tech Preview has now ended and is a fully fledged production supported feature. Notice also it’s been named ‘Flow Micro-segmentation’.
This feature is only available for AHV and once deployed will run for 60 days in ‘Evaluation Mode’, ideal if it’s something being considered for the data centre.
Scale-out Prism Central (General Availability)
Prism Central was previously a single VM deployment until the advent of AOS v5.5 when it became a 3-node deployment option, in Tech Preview mode anyway. Scaling out this management VM not only increases the total node/VM capability but boosts its high availability. The point of note here is the 3 Prism Central VMs will need to reside in the same location and not be split across multiple data centres.
The NX-6155-G5 model has been waiting for this release, support for 80TB (RAW) storage. A 2RU node with this capacity lends itself ideally for AFS as well as cluster expansions.
CALM (more functionality)
Following the release of CALM (Continuous Application Life-cycle Manager) new functionality has now been added in the form of :
- New target platforms, VMware ESXi (running on Nutanix), IBM Power (running on Nutanix) and Google Cloud Platform
- AHV / VMware ESXi chargeback
- Create an application by importing existing VMs (not initially deployed by CALM) and publish in Marketplace
V3 API (General Availability)
The approach of this release is to follow an ‘intentful’ philosophy, the idea being to move the programmatical elements away from the user and into the machine side.
Head over to http://developer.nutanix.com to learn about the APIs.
The Release Timeline page has been updated to reflect the above