Do believe the hype!
Many companies are migrating their data to the cloud. Cisco forecasts that by 2021, 94% of all workloads and compute instances will be processed by cloud data centers.
But Gartner’s 2018 hype cycle did have cloud migration falling into the “trough of disillusionment.” This could be because migrating data to the cloud isn’t easy. Organizations encounter numerous issues when trying to take advantage of the cloud’s benefits.
Organizations need to plan properly before undertaking such an endeavor. Let’s look at what you can do to minimize the issues that can plague a cloud data migration project.
Cloud Data Migration Defined
Cloud data migration is the process of moving an organization’s data from its current location to a cloud data center provider.
The current location of the data can be your own data center or another cloud provider. But the most common type of cloud data migration is moving from an on-premise data center to a public cloud. You can also migrate your data from one cloud provider to another. Yet a third type is going back from cloud to on-premise.
The Good Stuff
There are many reasons why organizations choose to migrate their data to the cloud. Here are just a few:
Reduced Costs
If managed properly, it should cost less to store and access your data in the cloud than in your own data center. You no longer own or rent the data center, and you don’t own the hardware. You’re renting storage space on someone else’s disk to keep your data.
You pay only for what you use, and the savings aren’t just in concrete financial costs, but also in time. No one needs to spend time ordering, receiving, racking, and stacking new servers. Ultimately, this saves you money.
Increased Productivity
With your data in the cloud, you can give your organization’s users access practically no matter where they are. They’d be able to collaborate on their work and projects without having to “go back to the office.”
Self-Service Deployments
With access to your cloud infrastructure and its data from almost anywhere, your team can do some things themselves. If they need to add a new database to a cluster of databases for data management, they can quickly spin one up. No ordering and waiting for delivery.
And this is only a sample of what you get with cloud data migration. What a time to be alive!
The Bad…
For all the good, migrating your data to the cloud comes with a lot of “extra baggage.” Many things can go wrong during and after a cloud data migration project if you’re not careful. Some organizations rush so much to take advantage of the benefits that little planning goes into the migration process. Because of this, you can encounter many challenges.
Limited Bandwidth Connection
Before you migrate your data, you need to look at your data center’s internet connection. The size of that connection will determine how fast your data can move to its new data center.
Are you moving terabytes, petabytes, or exabytes of data? Even with dual 100 Gbps connections, you’re looking at a half-day to move 1 petabyte of data. And that’s a best-case scenario!
So, if you don’t account for your internet connection, your data migration will take longer than expected.
Organizational Downtime
You likely have data stored in a primary location with a DR location, in case of failure. You have data redundancy—or at least you should.
When migrating this data to the cloud, you may not consider business continuity. While cloud providers would have us believe that failures don’t happen, you know they do. Earlier this year, Google Cloud was down with an outage. In 2017, AWS had a major outage caused by human error.
IT services fail. What will your organization do to access its data if your cloud providers have an outage? You need to plan for this for the business to continue operating.
Lack of Data Portability
Where and how your data is stored will determine how easy its migration is. Is your data portable? Can you migrate it without having to change its format? Is it stored in a silo, with another team restricting its access?
If your data is stored in an Oracle database, and your organization decides to move to MySQL, for example, what changes will you need to make? Are you locked into some proprietary license that restricts this move? Understanding your data’s portability will help avoid migration issues.
…And the Ugly
A fourth challenge that organizations face with cloud data migration is security.
CrowdStrike co-founder Dmitri Alperovitch is known for saying that “there are only two types of organizations: those that know that they’ve been hacked and those that don’t yet know.”
It’s a sad reality in the world today, but companies can get hacked even with the best efforts made to prevent it. But other companies seem less prepared, and that can be a challenge.
A cloud data migration project moves your data from a data center you likely own and control to one you have little to no control over. You rely on the cloud provider’s employees to be responsible and help you safeguard your data. But that’s not always the case.
You also need to be careful with migrating data stored in systems you may not fully understand. Or, migrating data that isn’t stored in encrypted form. An example is data found exposed in an open Elasticsearch server running on Google Cloud.
Mistakes like this can lead to very ugly things for your organization. You want to avoid them at all costs.
Cloud Data Migration Options
At its most basic level, the biggest challenge with cloud data migration is actually getting the data to the cloud. Like any IT process, challenges you encounter are sometimes made easier with the options you choose.
There are, broadly, three ways to get data from your data center to the cloud.
Online Over Public Internet
This option involves copying your data from your data center over your public internet to the cloud. This could be done many different ways, depending on the type of data you’re migrating. It could be as simple as a file copy, if your data is stored in files. Or it could be something more complex, like a database replication, if your data is stored in a database.
Online Over a Private Connection
If you don’t want your data going over the public internet, many cloud providers, like Google Cloud, offer private connectivity. This option is similar to the one above except you get a little more peace of mind, theoretically, of your data moving across a private connection to the cloud.
Offline With Provider Data Migration Services
Offline migration may be for you if you have so much data to migrate that online will take too long for your liking. Cloud providers like AWS have devices that physically transfer your data to them. Once all your data is copied onto their devices, it gets shipped back to them for transfer to the cloud.
Cloud Data Migration Tips
With all those challenges, here are some tips on how to benefit from cloud data migration and reduce the challenges.
- Assess initial infrastructure requirements for migrating data. Review your infrastructure before you migrate data to ensure that it can support data transfer within the time needed to migrate.
- Preserve the format data is currently stored in. Do your best to preserve the format the data is stored in today to save time of converting, and begin using data immediately after migration.
- Consider multicloud use for failover. You should implement a hybrid cloud if possible so users will always have access to their data.
- Involve your users. Migrating data to the cloud isn’t only a technological change; it can be a cultural one, too. Make sure to get buy-in from your data consumers early in the migration planning.
Bring Together and Conquer
Migrating data to the cloud can be a challenging undertaking. As you just read, there are a lot of things that can go wrong.
If you don’t take steps to plan and bring the right people together, you’ll fail. So plan for the impact to your infrastructure, team, and data users. The plan itself may not ensure success, but the planning process helps ensure reduced stress and total failure.
At that point, cloud data migration definitely won’t be hype in your organization. There’s no need to hype something that’s ordinary. And hey—being ordinary isn’t always so bad.
This post was written by Jean Tunis. Jean is the principal consultant and founder of RootPerformance, a performance engineering consultancy that helps technology operators minimize cost and lost productivity. He has worked in this space since 1999 with various companies, helping clients solve and plan for application and network performance issues.