In this blog, we’ll walk through our street proven approach to migrating analytic workloads to the cloud.
One of the biggest benefits of the cloud is that the barrier to entry is low because it’s very easy to spin up a server and start playing around. We recommend that you balance the ease with proper planning so that you can:
As with most analytics projects of any magnitude, we recommend that you start with a cloud readiness assessment to determine if you’re ready to make the move by reviewing your current environment – corporate and remote workforces, business needs, and priorities – your data, regulatory and security concerns, existing on-premises and cloud infrastructure, applications, and dependencies. With a firm understanding of your operations and existing assets, you’ll have a better idea of the best path forward for your cloud migration.
The steps of a cloud readiness assessment should include:
While cloud represents an opportunity for cost savings, if you’re not planning to decommission current resources, then you will be adding to your costs.
When it’s clear where you’re coming from, where you’re going, and who’s going to support you along the way, the next step is to develop a detailed plan. The cloud migration plan should contain the following elements.
Workload Assessment and Prioritization: Determine what can and cannot be migrated to the cloud. Many will start with a database platform migration which can be a “lift-and-shift” (moving an application or operation from one environment to another without stopping to redesign the app or operations workflow) or a migration from a legacy database to a cloud-native database. We’ve seen an interesting trend lately in that more companies are open to embracing open-source architectures and specifically open-source database engines like Postgres or MySQL to replace proprietary commercial offerings like Oracle and SQL Server and reduce costs.
From there, determine the mechanism of data ingestion. Will it come from an on-premises system or maybe a SaaS platform? This all factors into how much you need to invest in your network. Where you source your data from and what platform you’re pushing it into will guide your decision to leverage an existing on-prem ETL/ELT framework or migrate your data transformations to cloud-native or cloud-enabled tools or frameworks.
Even if you’re not moving your database or other source systems to your cloud environment, analytics platforms are another common first move into the cloud. In many cases you can start with a lift-and-shift to an IaaS provider. In other cases, you may consider a re-platform to SaaS offerings from the same vendor.
Don’t neglect embedded analytics or other integrations—such as dashboards in salesforce, SharePoint, mashup frameworks, or other third-party tools—or your authentication mechanism.
Platform Selection: Determine the appropriate platforms for your migration. It’s possible that different cloud models will be employed for different parts of the same overall workload. SaaS platforms, such as Power BI, Salesforce.com, Qlik, and Tableu, are turnkey offerings that provide you with an application as a service. You are completely insulated from and unburdened by the entirety of the tech stack behind the scenes. PaaS offerings, such as Snowflake, allow greater control and require additional management. IaaS offerings, such as AWS and Azure, are best equipped to support a broad variety of analytics workflows, but they also require a greater investment in the planning phases and the continued operational management. And while traditionally thought of as IaaS providers, AWS and Azure also have PaaS offerings (like Microsoft Azure SQL and Amazon RDS) and SaaS offerings (like Amazon QuickSite and PowerBI).
It’s not uncommon to leverage more than one of these options in a comprehensive cloud-based modern data architecture.
It’s not uncommon to leverage IaaS, PaaS, and SaaS platform options in a comprehensive cloud-based modern data architecture.
Macro Execution Plan – Big Bang, Phased, or Hybrid: With a big-bang approach, the project can be completed faster, but also with greater risk. A phased approach usually takes more time, but the migration team can focus on smaller chunks of the workload during each phase. If going this route, consider breaking it out by business unit or end-user application, and break up the complexity into separate phases to reduce risk in each phase. The hybrid approach often entails deploying the architecture to the cloud, but migrating content in phases, often aligned with business units or lines of business. Like a phased approach, you’ll probably run your legacy stack and cloud deployment in parallel until you can comfortably cut-over.
Micro Executions Plan: This is where you consider the individual needs of each tool or workload to determine the best micro-execution plan. A common industry term for these micro-execution approaches is the “6R”.
First, consider what is to be retired or retained as it currently exists. This enables you to immediately trim down your area of focus. The other four Rs of the “6R” cover the most common migration paths:
Execute Stage of a Cloud Migration
This diagram represents the general steps but will vary depending upon your migration strategy and architecture.
Once in production, keep optimizing your environment. The following are not one-time tasks; they should be performed continuously.
Automate shutdown of Dev/QA resources – doing this on weekends alone can immediately save you 30% with on-demand pricing. Shutting down overnight and on weekends can save up to 70%.
While the above steps are a great guide for your move to the cloud, remember that no migration will look the same. Your cloud migration process and overall strategy should be based on a long term direction that delivers real business value. Take the time to think about how cloud can be utilized to promote success at your organization.
We offer a free, one-hour cloud strategy session where you’ll get practical advice that you can use to get started or optimize your existing data and analytics workloads in the cloud.
To thrive with your data, your people, processes, and technology must all be data-focused. This may sound daunting, but we can help you get there. Sign up to meet with one of our analytics experts who will review your data struggles and help map out steps to achieve data-driven decision making.
In one hour, get practical advice that you can use to initiate or continue your move of data and analytics workloads to the cloud.
During your free one-hour cloud strategy session, we will: