Doing Converged Infrastructure Right: A Practical Approach

May 23, 2018
There are many use cases for converged infrastructure (CI) and hyper-converged infrastructure (HCI), but they can be frustrating to implement. Bill Kleyman offers tips for a successful deployment of CI and HCI solutions.

I’ve seen the power of converged infrastructure and I’ve experienced the frustration when something doesn’t go right.

There are so many great use cases for these technologies, including both converged infrastructure and hyperconverged infrastructure. They can save on space, optimize overall data center performance, introduce great new solutions like all-flash, modernize applications, improve density, and so much more.

Before we go too much further, let me define these two systems for the sake of clarity:

  • Converged Infrastructure (CI): With CI you see the integration of core resources and delivery technologies. The big premise behind CI is that it comes as a pre-validated, referenced architecture capable of being deployed in strategic data center building blocks. CI will bring together network, storage and compute resources in one integrated system. It allows you to remove data center resource silos and really begin to optimize virtual workloads. In these scenarios, management is done either at the hypervisor layer or at the CI management console or tool. A great example of this would be FlashStack, developed jointly by Cisco and Pure Storage.
  • Hyperconverged infrastructure (HCI): There are a number of similarities between HCI and CI environments. However, the biggest difference is how these environments are managed. In HCI, the management layer – storage, for example – is controlled at the virtual layer, as HCI incorporates a virtual appliance which runs within the cluster. This virtual controller runs on each node within the cluster to ensure better failover capabilities, resiliency, and uptime. In these types of models, you begin to see how technologies around software-defined storage (SDS) impact converged infrastructure systems. Examples of this would be Nutanix, Dell/EMC/VMware, and Scale Computing.

Here’s a visual guide from MicroAge:

Hyperconverged infrastructure has seen a five-year compound annual growth rate (CAGR) of 26.6 percent, and revenues are forecast to hit $7.15 billion in 2021, according to research firm IDC. In my own experience, I can tell you that we’ve been deploying the solutions quite a bit.

But what happens when you get it wrong? Maybe you forgot an infrastructure component, or underestimated your use-case. Or, maybe you’ve deployed the wrong system entirely.

To that extent, I can really narrow down the major challenges into five core deployment approaches and considerations.

Finding Success with Converged Infrastructure Deployment

Measure twice; cut once. When working with converged infrastructure, always make sure to size your environment properly. What are you deploying? Are these heavy applications or lightweight web portals? Are you supporting virtualization with high levels of user density, or are you running databases supporting business functions? Each scenario will dictate specific requirements.

When it comes to deploying a converged system, you’ll need to take a few things into consideration. It’ll be hard to cover all of the many scenarios out there, but I’ll try to give you an idea.

First of all, sizing is absolutely critical to HCI and performance. A huge benefit of HCI is the built-in capability around performance. This means that core components, like compute and storage, are already optimized for performance. This helps impact everything from Disk IO to CPU and memory. Some converged infrastructure systems come in pre-validated building blocks, making some of the sizing considerations much easier.

Consider the following five points:

  • Disk IO: This factor is absolutely critical to overall performance. If you’re working with database solutions, VDI (virtual desktop infrastructure) or even data analytics, working with Disk IO is so critical. Don’t deploy spinning disk where all-flash simply makes more sense, like VDI environments. This is why it’s critical to have not only the right type of configuration, but the right type of disk architecture in place as well. CI and HCI vendors can offer hybrid as well as hybrid solutions where SSD is thrown into the mix for better performance as well as all flash nodes.
  • Memory: Let me give you a specific example: SQL uses RAM to cache data pages. So, the more RAM, the better for performance. If you don’t get your design right, and you’re planning on hosting database systems, your applications will suffer and users will have a poor experience. I’ve seen this happen and it’s an issue you’d rather avoid. Outside of databases, many virtual environments are housed on some type of converged system, and some of these host virtual desktops and application. You’ll need to take into consideration load, user density, and the type of application or desktop that’s being delivered. If you know you’ll have CAD developers accessing a virtual desktop, plan for it by having enough memory available. On that note, you may even require integration with a GPU to help optimize memory utilization as well as graphics performance. Either way, good HCI designs need to come prepared with RAM capacity and scale to grow.
  • CPU: Let’s use the SQL example above. Index rebuilds and a heavy amount of seeks can cause CPU bottlenecks. SQL aside, this can also happen to hosted applications where too many users are trying to access resources. Saving dollars on CPU can absolutely impact you later. Make sure to plan and design CI or HCI platform so this won’t be an issue.
  • Network IO: The amount of data we’re creating within the data center continues to grow. Working with databases, applications, connected devices and business systems all place additional load on our infrastructure. The good news is that CI and HCI can absolutely make this easier. However, you have to design around your requirements. For example, in an CI configuration, you can select the type of network layer you require. Make sure you understand how much traffic you’ll be generating and plan around this need. The really cool part here is that both CI and HCI have direct integration with networking and switching capabilities, allowing you to easily create network segments where data interference is kept at a minimum. In an HCI design, when data is local to the application, residing on the same node, then you will get better network performance and efficiency. For CI architecture, you can integrate entire network solutions with SDN to simplify data flow and control.
  • Infrastructure Distribution: Take a moment and think about your business. Is it growing? Are you staying in one location or are you deploying smaller branch data centers or even edge locations? If you’re moving towards the edge, working with a node-based HCI platform can absolutely make sense. Although converged infrastructure is great, it can prove to be a bit much and bulky for highly dense and agile edge solutions. HCI, on the other hand, can offer some major benefits. When it comes to the context of edge, leaders in the space are making distributed IT as simple and affordable as possible. For example, Scale Computing allows you to deploy a single-node HCI architecture to support smaller offices and branch locations. In other cases, deploying a powerful Nutanix multi-node architecture (still pretty compact) might be required. The bottom line is that you need to understand the direction of your business to deploy the right architecture.

Finding Your Own Convergence Groove

There are a lot of great partners and vendor organizations who are ready and willing to help you out on your convergence journey. Before you reach out to anyone, take a moment to think about your business, your requirements, and your users. You may have legacy systems that you want to remove, or you may be deploying a new location. It’s so important to have a clear understanding of your use-case before you embark on deploying convergence for your organization. Do your best to leave no application or piece of infrastructure behind when doing your use-case research. It’s these little things that come back to bite you later.

My last piece of advice is to include business leaders in your planning. They can offer insight into future planning that’ll help shape your thinking. For example, I’ve been a part of projects where we were informed that new locations would be sprouting up, but they’d be small environments with limited space. This really helped us with design requirements to deploy the perfect solution. Business leaders are a part of the IT process and can really help make the ultimate design great.

About the Author

Bill Kleyman

Bill Kleyman is a veteran, enthusiastic technologist with experience in data center design, management and deployment. Bill is currently a freelance analyst, speaker, and author for some of our industry's leading publications.

Sponsored Recommendations

Get Utility Project Solutions

Lightweight, durable fiberglass conduit provides engineering benefits, performance and drives savings for successful utility project outcomes.

Guide to Environmental Sustainability Metrics for Data Centers

Unlock the power of Environmental, Social, and Governance (ESG) reporting in the data center industry with our comprehensive guide, proposing 28 key metrics across five categories...

The AI Disruption: Challenges and Guidance for Data Center Design

From large training clusters to small edge inference servers, AI is becoming a larger percentage of data center workloads. Learn more.

A better approach to boost data center capacity – Supply capacity agreements

Explore a transformative approach to data center capacity planning with insights on supply capacity agreements, addressing the impact of COVID-19, the AI race, and the evolving...

Photon photo/

In the Age of Data Centers, Our Connected Future Still Needs Edge Computing

David Wood, Senior Product Manager – Edge Computing at nVent, explains why edge computing will play a pivotal role in shaping the future of technology.

White Papers

Download the full report.

2021 Overview of the Phoenix Data Center Market

April 6, 2021
This report, in conjunction with NTT, continues Data Center Frontier’s market coverage of growing data center hubs. Explore further for a comprehensive overview of the Phoenix...