Enterprise Architect
Editor’s note: This is Part 3 of a multi-part blog series discussing the pros and cons of private cloud vs public cloud computing, and some specific technologies that can enable Private Clouds. The information presented is meant to be used for educational and planning purposes. Please contact Mainline for more information and assistance in building or progressing your cloud strategy.
Introduction
Occasionally, a bit of disruption is needed to shake up a stagnating operating model and help advance the art of the possible. Nutanix aimed to do this by simplifying the consumption and lifecycle of traditional datacenter infrastructure so much that it became invisible. This would allow IT practitioners to focus on higher order – and more relevant – business functions, which is still the holy grail of an IT operating model today. This is often referred to as “cloud.”
Nutanix
Founded in 2009, and shipping their first product in 2011, Nutanix had a philosophically simple goal to make infrastructure “invisible.” To do this, it had to effectively tear down traditional silos and replace them with something “software-defined.” We saw in the previous blog that VMware has a “software-defined datacenter,” but Nutanix pioneered this change. And the first thing Nutanix addressed with software was the storage layer.
By using software to pool local disks together into a unified storage fabric (called the Distributed Storage Fabric), it created an extensible, virtual array that wasn’t tied to specific hardware or hardware architectures. By itself, this is Software-defined Storage (SDS). When combined into a compute platform (like a standard x86-based server) with a hypervisor and local disks, it is known as hyperconverged infrastructure (HCI).
The tenants of HCI have always been:
• Simplifying procurement
• Simplifying operations
• Reducing costs
• Enabling (linear) scalability
One of the first hero use cases for HCI was in running Virtualized Desktop Infrastructure (VDI). Back in the day (when every year was the “year of the desktop”), much angst was produced by running VDI workloads on less-than-performant storage arrays. This was when all-flash arrays were just a glimmer on the horizon of affordability and most deployments had to make do with some sort of tiering solution. Nutanix’s default configuration was hybrid HCI, which made use of a small flash tier for non-sequential writes and as many reads as could be held in the hot cache. What made this HCI architecture particularly well-suited for VDI workloads was the distributed nature of the storage I/O, metadata, and flash. Nutanix often touted “data locality” (running the compute workload on the same node serving the storage), which was especially useful for systems that ran on 1 GbE networks, as those networks were never designed for storage workloads and could very quickly become the bottleneck.
Nutanix also offered an all-flash version, but even with the economics of NAND-based SSDs allowing for many workloads to run more cost-effectively for their higher performance gains, many customers still found the hybrid HCI architecture to mostly suit their needs. Obviously, today, all-flash economics make sense in many more situations, but there are still plenty of enterprise-grade hardware arrays that are both hybrid and performant.
But the real benefit of HCI was found in the unified management plane, automation, and data services that could be more easily instantiated and consumed. Such luxuries as a single, huge datastore with global compression and deduplication options (for all-flash) and native snapshot-based “backups” and replication allowed the VM admins to utilize lower-level storage more effectively. Plus, because the storage was designed for the singular use case of running VMs, administrators enjoyed nearly unparalleled visibility into the goings on of their infrastructure, something only really experienced by Tintri VMstore adopters at that time.
While the VM data was stored on a SDS array that Nutanix built and controlled, the VMs themselves could initially only be executed and managed in VMware’s vSphere (ESXi) and, later, Microsoft Hyper-V. Allowing a choice of hypervisor was a deliberate decision by Nutanix and was made possible by architecting their storage and management solution to be portable; it all ran within a shared-nothing clustered VM architecture (CVM), not within the kernel of an OS loaded on the physical server. This allowed for great flexibility and full hypervisor migration capabilities. This is one of the reasons why Nutanix was allowed at VMworld; they were seen as an enablement solution, not as a competitor.
But, in time, Nutanix developed their own KVM-based hypervisor, Acropolis, that would compete with the other two hypervisors they supported. It initially supported their ‘storage-only nodes’ as a way to get more storage into the cluster without having to scale up compute resources (or license their underlying 3rd party hypervisor). Because Acropolis Hypervisor (AHV) only had to support the CVM, it could be brought to market with a relatively sparse feature set. But new features, functions, and optimizations were slowly introduced that really put AHV in the running to replace either of the other two mainstream hypervisors.
And that’s exactly how Nutanix is positioning their platform today: a full replacement to traditional virtualization stacks and one that runs a lot more like a private hybrid cloud, with all the automation, observability, and data services built right in. This, then, is the value proposition of Nutanix: a true private cloud with all the mainstream features of a public cloud and competing directly with VMware’s VCF offering.
Unification
When you use a Nutanix platform, one thing stands out: unification. From the beginning, Nutanix has sought to integrate its varied capabilities as a singular entity. Contrast this with how other companies have done product/feature integration – I’ll pick on VMware here, as it’s the closest competitor for Nutanix and for no other reason – and you notice how seamless everything is. VMware built out the VCF stack from its varied product / business units, which, in turn, were formed mostly by acquisition. This is a challenge that has just recently been addressed and the fruitfulness still to be seen with the upcoming release of VCF 9. But Nutanix has had more than a decade of deep product and capability integration and professional UI/UX design to blend everything together, so it appears as a single system from which to enable many features.
Feature Sets
And those feature sets include most everything you could want in a modern hybrid cloud platform:
- Virtualized:
- Compute
- Storage
- Networking
- Disaster Recovery
- Security
- Cloud Native Services:
- Storage Services:
- File
- Object
- Containers
- Data Services for Kubernetes
- Database-as-a-Service (DBaaS)
- Enterprise AI
- A Cloud Control Plane for:
- Federated Management
- Cloud Management
- Self-service
- Lifecycle and Capacity Planning
- Cloud Costing / Compliance
- Governance
- Data Governance
-
- Third-party Integrations for:
- Backup
- EUC
- Security
- CI/CD
- AI/ML
- Third-party Integrations for:
In addition to all the on-premises features mentioned above, the entire stack also works in AWS and Azure on Bare Metal Instances. Called Nutanix Cloud Clusters (NC2), the two largest hyperscalers in the world can extend out on-prem capabilities at a global scale, facilitating global and regional deployments and BCDR imperatives. License portability allows for a flexible consumption, so you’re not locked into an architecture and can easily pivot to suite your technical and business needs.
Dell Partnership
And for all those who are sold on the idea of the benefits of a software-defined stack but are leery of exclusive hyperconverged storage on the compute platform, Nutanix is finally addressing those concerns with 3rd party external storage support. Starting with their partnership with Dell, the “Nutanix Cloud Platform for Dell PowerFlex” offering effectively decouples the compute and storage layers to be architecturally more like traditional 3-tier storage, but with the deep SDS integration you get with unified HCI management. The result is an independently scalable storage system that acts like a unified HCI system. Truly, the best of both worlds.
Modern Cloud Computing Platform
This is the stuff that makes up a modern cloud computing platform: flexibility, portability, and a unified administration experience. Nutanix has proven that they have great vision and ability to execute. Combined, these attributes make Nutanix an excellent choice as the foundation of your private cloud.
Hybrid Cloud Solutions from Mainline
Need help getting started or taking the next step to a modern cloud computing platform? Mainline’s experts help our clients craft migration and strategy roadmaps to ensure a return on investment. Our goal is to create a Hybrid Cloud infrastructure that offers scalability, agility, and cost-effectiveness to deliver optional business outcomes.
Our experts can assist you from the initial assessment through post-implementation support to selecting cloud services that can help you maximize and sustain business efficiencies. Contact your Mainline Account Representative directly or reach out to us here with any questions.
You may be interested in:
BLOG: The Case for Private Cloud – Part 1: A Winning Strategy
BLOG: The Case for Private Cloud – Part 2: VMware Innovation