Today at the .NEXT Conference, Nutanix unveiled the Xtreme Computing Platform (XCP) composed of two product families, Nutanix Acropolis and Nutanix Prism, XCP.
Prism is responsible for delivering convergence of storage, compute and virtualization resources into a unified system that provides end-to-end view of all workflows – something difficult to achieve with legacy three-tier architectures, while Acropolis builds on the core capabilities of the Nutanix hyperconverged platform and include all the low-level operations in the Distributed Storage Fabric, the App Mobility Fabric and the Acropolis hypervisor. Yes, Acropolis also includes a native hypervisor based on the proven Linux KVM hypervisor.
KVM an open source hypervisor that is gaining traction in the enterprise as a virtualization solution that provides high performance, scalability, and cost efficiency. First introduced into the Linux kernel in February 2007, it is now a very mature product and is probably the most widely deployed open source hypervisor.
KVM is fast gaining ground with backing from the likes of IBM, HP, Red Hat, BMC Software, Eucalyptus Systems, SUSE and Intel. They are educating users on best practice on deploying the hypervisor, largely to promote an open environment for competition against the likes of VMware and Microsoft.
KVM has all the security features that VMware has plus some more – and it has been awarded Common Criteria Certification at Evaluation Assurance Level 4+ (EAL4+). The Common Criteria is an internationally recognized set of standards used by the federal government and other organizations to assess the security and assurance of technology products.
Nutanix has always been upfront about being agnostic about hypervisors; no matter what hypervisor organization happen to choose (today NOS supports KVM, Hyper-V and ESX, more options possibly to come soon). I personally believe hypervisors will be the BIOS of the future. It is there, but nobody really pays attention to it. With KVM momentum Nutanix realized that whilst KVM is a battle tested hypervisor there was not a good enterprise-grade management story for the hypervisor, leaving many administrators to manage independent hosts; and this approach is clearly not scalable.
Since NOS 4.1 Nutanix has been quietly shipping a management platform that allow organization to execute virtualization management operations with KVM (manage virtual machines, live migration, networking etc.). This management platform follows Nutanix architecture, operating in a completely distributed architecture where there are no single points of choke or failure when managing virtual machines, servers, clusters, or datacenters. This control plane, also known as Acropolis, has been designed from the ground up to operate in a truly distributed fashion following web-scale principles like elastic services, granular scalable units, shared-nothing etc.
It is easier to demonstrate the architecture differences when comparing to something we are all familiar with; VMware vCenter. vCenter is not a scalable architecture and multiple vCenter servers are required to manage a large vSphere environment. Given the architecture constraints VMware created vCloud Director to serve as an overlay aggregator and at the same time simplify self-service. vCloud Director has since been converted into a Service Providers product and organizations still need to deal with multi-vCenter management. Furthermore, if a vCenter is down all server and virtual machines under management are not accessible for management purposes.
The Acropolis Limitless Scale
Acropolis is an overarching framework that provides a distributed management architecture with no single-point of failures, storage fabric, migration services, heterogeneous services and orchestration APIS that is able to scale limitlessly while maintaining management via single-pane of glass (PRISM) and single cluster approach.
…but I don’t know KVM and I am not interested in learning nor have time for it.
Infrastructure should be simple, it actually should be invisible. Nutanix engineering and design team have been able to completely abstract the complexity related to deploying the hypervisors, manage iSCSI LUNs, operate the CLI etc. All operating functions are available via the awesome HTML5 Nutanix PRISM interface where administrator operate storage, virtual machines and networking components with ease.
With the current release administrator are able to execute the following tasks using PRISM.
- VM Operations: Create, delete, update, power, pause/resume VMs, access via remote console
- Resource Scheduling: Determine initial placement of VMs based on CPU and memory availability
- Migration: Live migrate VMs across nodes
- Snapshots and Clones: Rapidly create efficient snapshots and clones of VMs
- High Availability: Auto-restart VMs on a different host upon host failures.
- IP Management: Assign dynamic IP addresses to a VM at creation for the life of the VM. Acropolis uses VXLAN tunnels for IP address management.
- Analytics: Monitor and report key metrics for both VMs and the infrastructure, including hypervisor and physical nodes
- Remote Console: Remote access all types of virtual machines and operating systems to operate at the console level.
- Host Profiles: Standardize configuration of hosts within a cluster based on user-provided information
- Virtual Networking: Setup and configure VLAN-backed virtual networks spanning the cluster. Use built-in IP management
- Upgrades: Non-Disruptive rolling Hypervisor upgrade in a Nutanix cluster
- Maintenance Mode: Put hosts into maintenance mode for upgrades and host removals
- Scaling: Easily add and remove nodes in minutes
The next release NOS (yet this year) will also include support for features like active resource scheduling (DRS), guest customization and other features I cannot disclose right now.
As always, all Nutanix features and properties can be fully managed via API using CLI, REST, PowerShell, Java Wrappers and Python. I will soon publish a new post discussing the API framework, but rest assured that everything can be automated in an Nutanix Acropolis environment. Nutanix already has a number of partners integrating their systems and creating workflows that add value to the solution.
Most Nutanix nodes today ship with vSphere and there are many influential reasons as to why organizations may choose Nutanix with vSphere or Hyper-V, including internal knowledge, brand trust, previous investments, ELAs etc. However, for certain workloads you may decide that the Acropolis hypervisor is a good option from a cost or licensing perspective; these workloads may include large analytics solutions like Splunk or development and testing environments.
A large number of enterprise customers are already using the Acropolis hypervisor in production today and have been actively providing feedback. This is not a 1.0 release, but rather a battle tested hypervisor with strong customer validation. Get ready and test it!
However, it goes without saying that having a hypervisor without applications certified to run on it doesn’t take us anywhere in the enterprise space. Acropolis hypervisor is currently undergoing numerous industry certifications, but makes me very product to introduce you to the initial ecosystem partners that include Microsoft Server OS’s, Citrix, Centos, Ubuntu, SuSE and many others.
How do I test Acropolis? Easy! The Nutanix Community Edition is only provided with Acropolis and it comes fully featured for up to 4 nodes. Just get your server (only 1 needed) ready and start testing it. Apply here to get access to the Nutanix CE.
This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net