How do you think about your next-generation infrastructure? Is it simple to deploy and scale? Does it provide clear migration paths that don’t require forklift upgrades and overnight data center work? Can it provide the flexibility to support applications in both on-premises and cloud environments?
One way to help define what you need in your infrastructure is to analyze how it will actually operate in your environment—from the design of the architecture (Day Zero), to deployment (Day One), to operations (Day Two).
This article examines key factors to consider at each stage of the process. For a visual representation of these same factors, please view our lightboard presentation, “Datacenter History from a Customer-Centric Perspective—the Pursuit of Simplicity with FlashStack.”
This is the time for careful and strategic analysis of what you need today and to speculate to the best of your ability about what you will need over the next several years. It can be challenging because the pace of innovation has accelerated rapidly with the growth of cloud and the expansion of technologies like artificial intelligence and machine learning, which often come from the business side as initiatives with an accelerated timeline.
However, the accelerated pace of innovation is here to stay, so in your Day Zero analysis, you should make sure your infrastructure gives you the ability to easily incorporate innovation without rearchitecting and to deploy best-of-breed products when they become available. Part of this is to ensure you can modernize without having to go through a forklift upgrade every few years.
To look at it another way, what happens if your Day Zero planning is wrong due to items outside of your control? What if the business tells you to double capacity next year vs. over three years? How painful will that be?
The primary Day Zero consideration is the architecture. In today’s environment, you want simplicity, flexibility, scalability and easy access to innovation. This requires a core alignment built around statelessness, which will give you ultimate flexibility where you can pay as you grow.
With a stateless design in which compute, storage and networking are loosely coupled, you can scale as needed, without being locked into overprovisioning certain resources as you would, for example, in a hyperconverged solution.
In addition to a stateless design, you want to look at what will be required for Day One and Day Two. Although deployment and operations represent the next two phases in IT modernization, it is important to take them into consideration when evaluating your Day Zero architectural requirements.
You can often get a taste for Day One (deployment) during Day Zero by asking for a demo or proof of concept. Often products that sound great on the whiteboard or in slides will not perform well in a demo or PoC. Knowing this sooner rather than later is critical.
Any solution you choose should be simple to deploy—virtually turnkey—and simple to operate, with centralized management that leverages high levels of automation and analytics.
This is the deployment phase. Hopefully, you’ve chosen an architectural model that will take you into the future and provide a simple path to upgrades and innovation.
Reference architectures, converged infrastructures and hyperconverged infrastructures have all been designed to make it easier for IT teams to deploy solutions faster and more efficiently than traditional architectural stacks that separate the management of compute, storage and network resources. You want the best of each of these models, without any tradeoffs.
At this stage, you are looking to get your applications up and running quickly and efficiently, with no downtime and with the necessary performance and capacity required for each workload. Key characteristics to look for include:
- Validated reference architectures that are documented, validated and tested for a broad range of application-centric designs.
- Fast and reliable deployment using high levels of automation, analytics and API-driven integration with external systems from cloud providers and other potential partners.
- An architectural design that offers disaggregated flexibility for easy expansion, giving you the ability to scale and grow based on the needs of your workloads and applications, not on the design of your infrastructure.
What’s interesting here is that the “Day One Tax” isn’t just on Day One but also Day 180 and Day 365—any time you majorly rearchitect or expand a system. Masking complex implementations with services or residencies is an approach that will slow innovation and add continuing cost over time.
This is the phase that is operations-focused or “Business as Usual.” For most business executives, they’d like to not even think about Day Two—they literally hire people to take care of Day Two.
Executives notice Day Two issues for two main reasons—instability and the inability to innovate.
If a critical service goes offline, the organization notices—and not in a good way. Similarly, if an infrastructure requires so much day-to-day “care and feeding” that there’s no budget or time for innovation, IT becomes known as the “Department of No.”
What you don’t want at this stage is an infrastructure that forces you to do forklift upgrades. Otherwise, you will constantly be going back to Day Zero and Day One and repeating the cycle over and over again. The investment you make now should last a long time as the foundation for your on-premises infrastructure.
For Day Two, key factors include:
- A single-pane-of-glass management platform.
- Automated processes for day-to-day management and scaling.
- No need for forklift upgrades.
- Architectural design around performance and reliability.
- A single source for customer support.
- Flexible deployment options, including reference architectures, managed services options and loosely coupled infrastructure for independent scaling.
Today’s infrastructure has to be the foundation for your future. You have to think beyond the initial implementation and make sure you have an architectural model that is easy to deploy, manage and scale, with the least number of restrictions on when and how you can upgrade and modernize.
Starting with a stateless architecture is key, but it is also important to look at the other factors discussed in this article, including automation, centralized management, simplified upgrades and reliable, consistent support from your providers that is easy to access.
For more information on how you can view infrastructure modernization through the prism of Day Zero, Day One and Day Two, please review the resources and articles in this section and visit the video link mentioned above.