Nutanix & VMTurbo: A Match Made In Hyper-Converged Heaven?
AUGUST 10TH, 2015 by Joe Burke
Hyper-Convergence is all the rage these days. Just like “Software-Defined X” has steadily grown from a fancy marketing term to a reality, so has Hyper-Convergence, and for good reason. Hyper-converged environments aid in highly efficient scalability coupled with the elimination of infrastructure silos. It refers to an infrastructure that is largely software-defined with tightly integrated compute, storage, networking and virtualization resources. Rather than traditional infrastructure where each resource is typically handled by a discrete component that serves a singular purpose, you can think of this as what many call a “PO in a box.”
Many would argue that Hyperconvergence is the embodiment of the software-defined data center.
However, in order to be the most effective, the components must be integrated and not just simply bundled together. One of the major players in this realm is Nutanix. In the case of Nutanix, they converge compute and storage to form a single node used in their appliance. According to Nutanix, benefits include single unit to scale, localized I/O for high performance, software-defined intelligence, and distributed autonomous systems.
Phew. Okay, so let’s break this down in laymen’s terms. Use of cheaper commodity hardware, enhanced agility, improved scalability and efficiency, centralized systems and management, and did I mention lower costs? Sign me up!
How does it work?
Before we go any further, let me do my best (Nutanix experts feel free to slaughter me in the comment section) to explain what’s going on here.
Each node runs some sort of an industry-standard hypervisor. Think ESXi, Hyper-V, or KVM. In addition, each come equipped with a Nutanix Controller VM or CVM for short. The Nutanix CVM is what runs the Nutanix software and serves all of the I/O operations for the hypervisor and VMs running on that host. For the Nutanix units running VMWare, the SCSI controller, which manages the SSD and HDD devices, is passed to the CVM. The controller VM is where the vast majority of the Nutanix software and logic sits.
There’s also some other interesting things that are occurring on the backend to aid in performance and availability, and this is where VMTurbo comes in to the equation. Together, a group of Nutanix nodes forms a distributed platform called the Acropolis Distributed Storage Fabric (DSF). DSF appears to the hypervisor like any centralized storage array, however all of the I/O operations are handled locally to provide the highest performance.
Stay with me, we’re getting to where VMTurbo aligns in just a second! Quickly, let’s explore what’s occurring with DSF and the native Cluster-wide Tier balancing Nutanix is delivering:
“When the overall tier utilization breaches a specific threshold [curator_tier_usage_ilm_threshold_percent (Default=75)] where DSF ILM(Information Lifecycle Management) will kick in and as part of a Curator job will down-migrate data from the SSD tier to the HDD tier. This will bring utilization within the threshold mentioned above or free up space by the following amount.”
In other words, as hot data becomes cold over time, push it to HDD or lower tier storage to free up availability on SSD or higher tier storage to meet performance demands of the systems that require it. Awesome! So, how can VMTurbo aid even further in conjunction with what Nutanix is already offering natively?
Introduce a New Commodity!
If you’ve ever come across a VMTurbo blog at all, then chances are you’re familiar with “the market” and how actionable decisions to assure performance and control the environment in a desired state are derived. When it comes to Nutanix, we simply add another commodity help fine tune the desired state which in this case we represent as “HotStorage.”
First, we represent the various aspects of Nutanix underlying architecture in VMTurbo’s abstraction model:
Then, we introduce the new commodity to the existing market:
From here, we can build upon the existing performance related capabilities that Nutanix offers, and introduce Nutanix nodes as just another entity in the environment to control every facet of the virtual data center in healthy state by delivering VM placement decisions based on the combination of compute/network resources and best access to hot-tier based on hot-data demand. In addition, VMTurbo can now aid in questions surrounding when to scale up/down hot-tier storage driven by demand by adding more hosts to the cluster as systems grow.
Controlling and assuring the performance of any workload, on any type of infrastructure, anywhere, any time.
Kaynak : http://vmturbo.com/blog/nutanix-vmturbo-a-match-made-in-hyperconverged-heaven/?utm_source=linkedin&utm_medium=social&utm_campaign=resource-corporateblog&utm_content=nutanix-vmturbo-a-match-made-in-hyperconverged-heaven