LinkSource Blog

mac-computer-on-desk

How to Write Your Own Cloud Workload Solution In 3 Steps

If you missed our first piece called, “Let Your Workloads Guide Your Cloud Strategy”, go here.

It can be hard to determine which Cloud workloads are best suited for your Cloud migration. There are many variables to consider, like how mission-critical each workload is, or the systems and data dependencies of your applications.

To make the right decision, you need to be as thorough and clinical as a doctor examining a patient. Follow these three steps to create a Cloud strategy that treats and prevents your particular IT ailments:

Step 1: Diagnose Your Cloud Workloads

Understanding the characteristics of your workloads is extremely important to the design of an effective Cloud architecture. The more accurately your diagnose your workloads, the better your chances at improving overall performance with a Cloud workload migration.

Start by evaluating and categorizing your workloads. You may already have profiles of each workload that describe its purpose, the software and services running on it, and the resources it requires. This profile will help you categorize which workloads are mission-critical, which require the most protection and availability, and, ultimately, which are ready for the Cloud.

Your “diagnosis” of each workload should be based on:

Lifecycle: Is this workload cyclical, seasonal, or a one-off need?

Each workload is different, but they all go through similar “lifecycle” events, such as architecture, programming, testing, maintenance, integration, and management. Applications and software go through changes as they are provisioned, secured, migrated, reconfigured, and eventually retired.

Managing concurrent workload lifecycles can be challenging because you need to orchestrate multiple, fluctuating resource and availability requirements while maximizing efficiency and minimizing cost. Perform a thorough evaluation of each workload lifecycle, and look for patterns, redundancies, and any upcoming maintenance, upgrades, or configurations that may affect a Cloud workload migration.

Present and future goals: How do you want this Cloud workload to serve your business?

There are two different kinds of goals to think about when diagnosing each workload:

  1. Performance goals like increasing agility, improving service level, automation, and expanding storage volume and throughput, to name just a few.
  2. Business goals like improving operational efficiency, simplifying infrastructure, reducing IT costs, gaining customer insights, increasing productivity, and generally becoming a lean, mean, innovation machine.

Performance goals will help you design a Cloud solution that optimizes all your IT assets and resources. Business goals will help you make sure that the best of those assets and resources are going towards what matters most.

If you can’t clearly link a workload to a specific business goal, then it’s time to reconsider that workload. If you notice a particular workload directly supports multiple business goals, that’s a clue that you may want to prioritize that workload in your Cloud workload migration.

Strategic value: Is the workload critical to doing business or a key source of differentiation?

Diagnosing strategic value, though similar to identifying the business goals of each workload, is a distinct (and extremely important) step toward designing your Cloud strategy.

It’s not about how you hope to improve a workload; it’s about how that workload impacts business growth. Does it enable your employees to do great work? Does it contribute to customer experience and satisfaction? Would your business drop dead tomorrow without it?

A Cloud workload has strategic value if it is core to your business (meaning you can’t function without it). For example, Web and mail servers are pretty obviously core to most enterprises. Or, its strategic value may come from how the application, software, or task the workload supports provides competitive differentiation. A custom, client-facing application, or advanced data analytics platform, may be competitive differentiators for your brand even if they aren’t absolutely essential.

Step 2: Prescribe a Cloud Type For Each Workload

Take your list of Cloud-ready workloads, and determine which type of Cloud is the best fit for each. The best fit is the Cloud that’s most efficient and cost-effective, and also meets both the current and projected resource, agility, and scaling needs of the workload.

Assign each workload to one the three main Cloud types:

1. Private Cloud (Dedicated Servers/Dedicated Resources)

Whether on-premise or managed by a Cloud provider, the dedicated, secure, and more tightly controlled nature of Private Cloud makes it well suited for workloads that require carefully controlled environments, or performance guarantees. Some databases, legacy applications, and workloads with specific hardware dependencies are best run in a Private Cloud.

2. Public Cloud

Though you have lots of choice amongst managed third-party Cloud providers, AWS, Microsoft Azure, and Google dominate your Public Cloud options.

Public Cloud is characterized by vast resources and on-demand scalability in a shared environment with less control over security and performance. It’s a popular choice for application development and testing, Big Data analytics, and non-critical storage.

With Public Cloud solutions, it’s also up to you to understand your peak and off-peak use times and optimize your Cloud to suit. Perhaps your apps are in high-demand during business hours, but used very little at night. Or maybe you need 15 instances during weekdays, but only a few on the weekends. When using a Public Cloud the responsibility is on you to adjust your Cloud to accommodate these fluctuations in workload use.

3. Hybrid Cloud

Though Hybrid is an annoyingly vague term that can mean any combination of Private and Public, on-premise and managed Cloud solutions, it’s still a useful category when creating your Cloud strategy. You may find that lots of your workloads need a Hybrid Cloud solution because they need both security and agility. Hybrid is also a good fit for workloads that only need occasional or temporary boosts in capacity.

Now that you know what Cloud type each workload needs, you can step back and get a picture of what your ideal Cloud mix may look like.

Step 3: Make a Cloud Workload Migration Strategy

Now you’re ready to prioritize which workloads to move to the Cloud, and plan exactly when and how to do so. Your Cloud Workload Migration Strategy should take into consideration everything you’ve just learned about your workloads, including their:

  • Resource needs and dependencies
  • Business and performance goals
  • Strategic value
  • Ideal Cloud environments

If there is a high concern for data security and administrative control of your most critical workloads, your plan may include first building a Private Cloud or turning to a dedicated Cloud service provider, trying out Public Cloud bursting with less critical workloads, and gradually transitioning to Managed Public Cloud services where you can take advantage of advanced automation and expert monitoring.

But that’s just one example of the many different paths you can take to the Cloud.

To help you prioritize workloads for the Cloud and define your short and long-term Cloud Workload Migration Strategy, answer the following questions:

  • Are there smaller or less critical workloads we can first use to gain experience in the Cloud (like training wheels)?
  • Do we have development and test workloads that would benefit from a surge in resources?
  • Can we benefit from offloading some non-core workloads to the Cloud, and focusing our internal IT resources on workloads with the most strategic value?
  • Will our more sensitive workloads be most secure on-premise or under the eye of a dedicated MSP?
  • When do we need to be prepared to handle traffic and resource spikes? Is there a high traffic eCommerce season we want to be ready for, or a new application we’d like to roll out soon?
  • Can any workloads be divided, placing some components in the Cloud and keeping some in house?
  • Do we have customer facing applications? What will they benefit the most from right now: more computing resources for faster performance, 99.999% guaranteed uptime, or tighter security and attack prevention?

This list of questions is nowhere near comprehensive, but it should get you thinking from all angles as you define your Cloud strategy and start making moves to the Cloud.

Want to learn more about how Cloud migration can help your workloads? Read our blog, “Let Your Workloads Guide Your Cloud Strategy”

eBook: Outsourcing Communications Technology Lifecycle Management