Share this post

May 18, 2015

To Converged Infrastructure and Beyond

Bill Kleyman - MTM Technologies

Avatar
Modern organizations are being forced to reassess how they provide resources and support applications in the face of shifting user demands. Traditional data center technologies are coming under increasing scrutiny as more and more customers start to explore alternatives to the compute, storage and networking technologies they deploy today.
 
Organizations of all shapes and sizes are considering a converged and unified platform as they seek to deliver more resources with fewer moving parts to increasing numbers of users. But there are different phases of convergence and it's worth looking at some of them in closer detail.
 
First steps to convergence
 
Let’s start with storage. Converged storage removes a lot of the moving parts associated with traditional storage by consolidating components into a highly redundant node-based platform built around capacity, performance and optimization. It has many use cases. For example, organizations can offload workloads to a converged storage system instead of buying another rack of disks when deploying storage for virtual desktops. This makes it less expensive and easier to manage with better integration and the capability to meet future business needs.
 
Another approach is to deploy a unified architecture which directly connects network, storage and compute through a fabric backplane. This is a good solution for larger enterprises or data center providers because it can scale many chassis across a number of locations and works well for deploying a platform standardized on a single vendor across a number of hardware technologies. While it delivers on policy control, workload delivery and security, a unified architecture can be more expensive than some of today’s newer solutions.
 
So what's new?
 
One of the most powerful new solutions is a converged infrastructure which combines storage and compute in a single physical appliance. While networking is integrated to some extent, it's typically though a hypervisor.
 
Unlike a unified architecture which focuses on high end compute, storage and networking technologies for the data center, a converged infrastructure is small, powerful and dynamic in use cases. By integrating hypervisor technologies and abstraction services with the underlying physical architecture, converged infrastructures also introduce innovative software-defined technologies.
 
Take an organization opening a new branch location with a few hundred users that needs to have locally accessible resources for virtual applications and a few hosted desktops. A multi-node converged system would take less time to deploy than a traditional server and storage architecture, cost less and simplify management by unifying data center controls.
 
Hold on, there's more to come…
 
But this wouldn't be the technology industry we’ve come to know and love if a converged infrastructure was the end of the story. Of course, there's more: hyper (ultra) converged infrastructure. Like a converged infrastructure, hyper-convergence employs a software-based approach but takes it radically further by using software to control all resources living on top of the hypervisor. Hyper (ultra) converged systems can be deployed on any piece of hardware and any kind of hypervisor using APIs and VM translation technologies. This feedom of choice gives organizations the opportunity to build their own converged infrastructure platform by deploying the same software to manage storage and compute.
 
Convergence software is also intelligent enough to translate the VMs from one hypervisor to another if a business has a second location running a different hypervisor. Hyper-convergence can be deployed on whatever hardware delivers the most efficient use case so long as it meets the underlying requirements. Enterprises – large, medium or small – can extend the data cetner beyond their private architecture by using APIs to integrate with public cloud components and other hypervisors. They can also develop a next-generation data center and cloud model by integrating automation and orchestration controls.
 
The cloud's the limit
 
Today’s smart enterprises segment entire workloads on their own physical infrastructure or extend workloads cross-hypervisor while creating an agile cloud environment.
 
There will be even more creativity on both the hardware and software sides of storage in the future as the data center and cloud model evolves, leading to better integration of resources with powerful APIs and the interconnection of cloud components with private, public and hybrid cloud platforms. Organizations need to recognise that things have moved well beyond traditional data center server, compute and storage platforms. They need to hop on hyper-converged wave
12345
Current rating: 5 (3 ratings)