'Hyperconvergence' - What Cisco learned from the industry rush to 1 dot 'Oh' (Part 1
- Feb 15, 2017
- 5 min read
empowering the technology luminary to re-define 'simple' and re-invent 'complete.' - Part 1
The worst possible outcome of solving any problem is doing so at the expense of introducing a new one.
I like to refer to the pitfalls of first-generation Hyperconvergence, or 1 dot ‘Oh”, as in ‘Oh I didn’t realize we couldn’t do that,’ or:
Proof by Contradiction
In logic, proof by contradiction is a form of proof, and more specifically a form of indirect proof, that establishes the truth or validity of a proposition. It starts by positing the opposite proposition is true, and then shows that such an assumption leads to a contradiction.
I’m certain that if you’re a mathematician reading this you’ll be quick to point out that my use of this logic to tell the story of how everyone in the industry blew it when they rushed to be first to market with a Hyper-converged infrastructure (HCI) appliance is not entirely appropriate. I’m in no position to argue otherwise – mathematical logic is not my strong point and I'm not entirely sure if a hyphen is required at all times between Hyper and Converged or just some of the time. For the sake of this post, we’ll use the hyphen.
Before I get into the problems that first generation HCI introduced let’s talk about the Data Center, the brain of a large company, focusing on 5 big animal factors that led to the emergence of a $2 Billion market in 2016 alone.
The Data Center
The Data Center is a properly outfitted room with a large amount of equipment associated with supplying power and cooling, and often automatic fire extinguishing systems. It’s a place where the most critical business processes are run on sophisticated computer hardware hosting a myriad of software applications and operations that turn the cogs of profit and revenue, optimally resulting in growth and success for it’s owners, employees, partners, and shareholders. The Data Center is the operational responsibility of a company’s Information Technology (IT) department headed by the CIO (Chief Information Officer.) In many cases, the CIO reports to the CEO of the company but commonly they also report to the CFO. By placing the IT department under finance, the message is loud and clear: Technology is purely an operational concern, and the main focus of technology is cost cutting.”
Business-IT Alignment
Business to IT alignment is the correspondence between the business objectives and the IT requirements of an enterprise. These two factors, while possibly contradictory in nature, more so than ever in the modern era of technology-driven change management and it’s resulting challenges, require an alignment that has to be maintained over time in order to ensure the success of an enterprise.
Operational Business-IT Alignment
Operational Alignment is one key area of Business-IT alignment. Central to this element is the IT group’s adoption of an operating model for delivering services and support that meshes with the way the company works as a whole. Simply stated, if a strategy-driven alignment (that IT projects and budget can be directly tied to the company strategy) is about what is getting done, then an operational alignment is all about how it gets done. In particular, it's how IT services are delivered.
Enter ‘The Cloud’
As companies rapidly adopted cloud computing (the use of a network of remote servers hosted on the Internet to store, manage, and process data, rather than a local server in the data center) they were introduced to new IT consumption models. Fast and easy is the underlying theme to adding compute capacity in the cloud and this new IT consumption model introduced the need for a new operational model when company CFO’s began to expect a cloud economics model for their on-premise (hardware they own in the data center) infrastructure needs.
Cloud customers enjoy a pay as you go, scale on demand economic model that responds adequately to their business needs. These can be the kind of needs that reflect a sense of urgency where projects need to move quickly or when business is slow in down cycles there needs to be an appropriate response as well.
This resulted in a new customer demand of it’s IT vendors - “Don’t make me buy Data Center from you in big chunks. Allow me to buy it in a granular method by allowing me to flex it up and flex it down.” This is what Cisco refers to as:
IT moving at the speed of business.
Virtualization
The fundamental technology that powers cloud computing is virtualization. Virtualization is a software that separates physical infrastructures to create various dedicated resources. Virtualization software makes it possible to run multiple operating systems and multiple applications on the same server at the same time.
Cloud computing is the delivery of shared computing resources, software or data — as a service and on-demand through the Internet. The cloud often includes virtualization products to deliver the compute service.
Putting it all together > As companies attempt to operationally align their Business to IT, with a renewed focus on how their IT services are delivered, they now expect to leverage the benefits of cloud computing's IT consumption model, and it's cost effectiveness, to move their IT at the same speed as their business; transforming the architectural complexity of their Data Center into a more flexible, and responsive operation.
Enter the demand for Hyper-converged Infrastructure
To properly understand hyper-convergence you need to take a step back to it's IT predecessor - converged infrastructures (CI), the rack-based packages that combine existing storage, server and networking components. These systems provide either a complete storage and compute infrastructure at the rack level by selling commercially available components bundled as a single SKU or a reference architecture from which users or integrators can assemble a complete system such as the Cisco/NetApp FlexPods.
What a customer primarily gains from CI is a significant reduction in deployment time and a simple, common management tool for all their physical and virtual server components.
[What's an Appliance?]
While CI systems bundle rack-level components into an integrated easy to manage offering, a Hyper-converged appliance (HCA) combines the compute, storage and networking functions into the same chassis. To be classified as an HCA, a product must meet the following criteria:
[Hardware and software] While software is always at the core, HCAs aren't software-only products. Some first generation HCI offerings only delivered a software defined storage (SDS) application.
[Single-vendor product] An HCA should have an orderable part number and support for an end-to-end platform solution.
[Comprehensive management] An HCA is a turnkey product that, at a minimum, manages and simplifies the setup and initial configuration of the product and, ideally, provides simpler operation.
[Storage federation.] There must be a software layer that virtualizes physical storage capacity in each node and federates those nodes into a common management point.
[Hypervisor] HCAs include an embedded hypervisor for running compute workloads and, typically, to run the virtual machine (VM) based management and storage software.
With that, most HCA vendors rushed to market and focused only on delivering simplicity. In doing so they made some tradeoffs that limited the potential of Hyper-converged infrastructure:
Easy, building block scaling often came at the expense of inefficient, lock-step ratios of compute and storage (ie, you must buy the appliances in groups of 3 fixed model configurations) /THIS IS NOT SIMPLE/
Software defined storage stacks built on conventional, write-in-place file systems, show limitations in performance and data optimization. /THIS IS NOT COMPLETE/
Rather than integrating with existing tools, and offering the flexibility of extending your storage and compute capacity needs into your legacy environment new silos of management and policies were introduced instead. /THIS IS BAD/
Most importantly, in the context of Hyper-converged advantages of offering a distributed, clustered, fully integrated and scalable system: networking was never fully integrated with compute in these products. /THIS IS SAD/
In Part 2 we'll analyze the wait and learn approach that Cisco utilized to deliver a truly simple and complete Hyper-converged offering that renders inferior competitive offerings as obsolete.
Comments