Matt Hilbert

21 July 2023

Managing the challenge of migrating to the cloud

Migrating to the cloud has many advantages … until you’re the one tasked with making it happen. For many data professionals, this will be a first step from the traditional, predictable challenges of managing on-premises servers to the many and varied databases and platforms in the cloud.

As many have discovered, it’s not simply a case of ‘lift and shift’. There are a lot of considerations in terms of which flavor of cloud to choose, what data to move to the cloud, how to move it, and how to monitor it once it’s there.

The key is to break it down into four steps, not one.

Step 1: Get to know the cloud

The cloud isn’t one cloud, it’s many clouds, with Flexera’s 2023 State of the Cloud Report showing that 58% of organizations are using multiple public clouds. A deeper dive reveals that enterprises with more complex environments have 50% of their workloads and 48% of their data in public clouds. SMBs, meanwhile, have 67% of their workloads in the public cloud and 63% of their data.

As might be expected, the leading public cloud provider is AWS, with Azure following close behind, Google a distant third, and Oracle and IBM tailing in their wake.

The variety of cloud offerings in use is actually good news because every player has a vested interest in attracting business to its own cloud. As a result, they all offer a wealth of learning content because they want to make migrating to their service as easy as possible. That said, learn as much as you can using resources like Getting started with Amazon RDS, Azure SQL documentation and Cloud SQL on Google Cloud.

The objective here is to get a broad understanding of the different cloud offerings so that you can decide which is the best one for your organization’s needs. For example, AWS has a vast toolset and proven capabilities, but it has a complex pricing structure and there is a steep learning curve. Azure is tightly integrated with Microsoft products, and it’s good for hybrid clouds, although some services are more expensive than other vendors and others have limited availability. Google is good for data analytics use and machine learning, but it has fewer features and is not focused on enterprises.

Step 2: Plan what you’re migrating to the cloud

With an understanding of the cloud offerings and probably an idea of which one you prefer, it’s time to look at your existing server estate and start to make some decisions. It’s unlikely you’ll want to migrate every instance and database, for example, so you’ll need to agree which ones are the best candidates.

It’s a good idea to bring in a monitoring tool here if you don’t have one because it will end up saving a lot of time and money. The reason is baselines and trends. What are your computing needs? How do your servers perform? When are the peaks and troughs and the performance bottlenecks? Where are there pressures on resources or excess resources?

Another good move, and one which will help you decide which data to migrate first, is to look for common database issues that might cause problems. Typically, these would be outdated syntax, legacy views, functions and stored procedures, linked server queries and synonyms.

With the answers to issues like those, you can optimize the databases before you migrate them to get the best value out of the cloud. Remember, you can change your compute needs over time very quickly with the cloud without buying a new storage area network (SAN) or server.

Step 3: Set up a Proof of Concept

Before you migrate anything to the cloud, you’ll want to know how it will perform in the cloud. With a PoC, you and your team can get familiar with the new approach and start experimenting. It will also help you discover if the cloud offering you’ve chosen does live up to the promise and deliver on real-life issues.

This is a good time to test out your monitoring solution as well. Can it handle monitoring in the cloud alongside your on-premises monitoring, what kind of metrics does it provide, and does it give you the insights you need to maintain performance?

Step 4: Be proactive, not reactive

Once you’re ready to make your first migration, make sure you’re prepared if things do go wrong. You’ll need to keep an eye on performance issues like latency, interoperability, dependencies on no-cloud apps and unexpected downtime.

A good monitoring tool comes in again here because it will help you be more proactive and anticipate issues before they materialize. And remember to have contingency and recovery plans in place just in case.

Migrating to the cloud summary

Migrating to the cloud is a challenge for every organization, so remember you’re not alone in terms of the issues and questions that will come up. It’s about choosing the right cloud provider first, and there are several options, each of which offers different advantages. Then there’s the question of which workloads and data you want to move to the cloud, and why. And finally, there’s being able to monitor what happens when you’ve moved to the cloud, and being able to compare the performance with established baselines to ensure you’re making the most of it.

A monitoring tool like SQL Monitor has a place here because it lets you monitor your entire database estate, whether on-premises, on Virtual Machines, or in different flavors of the cloud, from one screen at one time. Before you even make the move, it can also help you identify existing issues with a database that might cause problems, and give you an accurate picture of the I/O performance and help you optimize it to make the cloud as cost-efficient as possible.

• To discover why SQL Monitor is the tool of choice for monitoring cloud databases, visit our cloud resources page.
• To find out more about SQL offerings in the Azure environment, watch this free on-demand webinar hosted by Pamela Mooney: How to Size & Price a Cloud Environment – A Guide for the DBA.
• For a non-technical guide, read 7 Essential Factors for a Successful Cloud Migration.
• And to find a wealth of helpful resources, refer to Redgate and the Cloud: How we can help you.