Many things have driven the rise of the cloud data warehouse. The cloud can deliver myriad benefits to data teams, including agility, innovation, and security. With a cloud environment, departments can adopt new capabilities and speed up time to value. More users can access, query, and learn from data, contributing to a greater body of knowledge for the organization. What’s more, the rapid shift to remote work has enhanced the benefits that moving to the cloud unlocks. For businesses focused on cloud data migration, one question remains: How do you get there?
As the race to the cloud data warehouse has unfolded, one thing has become clear: Simply lifting and shifting data does not achieve business objectives in a timely fashion. For this reason, cloud migration has been a growing priority for organizations in recent years.
Yet cloud data migration is not a one-size-fits-all process. It’s becoming increasingly important for organizations to plan out their migration to ensure compliance and security are maintained throughout – and costs are kept in check. With this in mind, there are a number of best practices to consider as your organization begins to move to the cloud.
What is On-Premises to Cloud Migration?
Cloud migration is the process of moving an organization’s data, workloads, apps, and other digital assets and IT services into the cloud from an on-premises infrastructure.
Not every on-premises to cloud migration is going to look the same because companies have different starting points. That said, there are two general models a company follows:
For data leaders moving an entirely on-premises environment to the cloud, you might use an offline migration. This involves transferring the digital assets stored on physical storage media to the cloud. This method is often time-consuming and expensive. Organizations with large volumes of data may consider this option, though it’s given way to more modern methods in recent years.
An online migration is less time-consuming and expensive because you’re not using a portable device to transfer the data. With an online migration, you copy your data from the on-premises infrastructure to the cloud using the cloud provider’s direct connection or the public internet. Most of your costs come from network bandwidth and data ingress fees.
What are the Benefits of an On-Premises to Cloud Migration?
Most companies know that the cloud offers plenty of benefits. But to make the business case, you’d be well advised to go beyond high-level “volume and velocity” conversations familiar to IT professionals. Today, the cloud can be used to mitigate risk, reduce costs, and accelerate efficiencies.
If you’ve got people working from anywhere, having applications and data in the cloud means you have limitless accessibility. Even small companies can be global companies today. The cloud gives you the global accessibility that removes bottlenecks for workforce members. Everyone has the same access to marketplace applications and advanced technologies.
By migrating to the cloud, you also create a more reliable and resilient infrastructure. Cloud providers have the money to invest in resources, like data backups, data rendunant storage, and backup power supplies. This means that you don’t have to worry about business interruption from a weather event causing a power outage or a physical machine dying.
The cloud is cost-effective because you only pay for what you use, like computing, network, and storage, which means you can more easily scale up or down as needed. A successful migration can also help reduce license costs for on-premises capacity, as well as reduce storage costs by cutting down on data redundancy.
Migrating to the cloud means your IT team doesn’t have to take care of the tasks that the cloud provider manages, like upkeep and maintenance.
All of these benefits culminate with having a centralized location for security. You’re still going to be responsible for managing the security in the cloud, but your service provider has tools that can help you. Authentication, authorization, logging, and auditing are part of your cloud platform, so you have a single location to streamline all security activities.
What are the Types of Cloud Migration Techniques?
Once you’ve built the business case and gotten your cloud migration approved, you need to figure out how to move your on-premises infrastructure. Usually, companies choose from one of three techniques.
Lift and Shift
In this model, you take your on-premises digital assets and just move them to the cloud “as-is.” This means that you’re not making any changes to the cloud servers that are running the applications. CIO Isaac Sacolick argues that organizations should avoid lift and shift migrations. “They should define the problem, find the value, and architect a solution that meets objectives. Sadly, lift/shift often leads to more than two migrations.”
Improve and Move
With this model, you do some pre-work to optimize your digital assets. This process means that you re-architect your on-premises digital assets first and then move them to the cloud. For example, you might want to update some applications by introducing scaling or automation without completely repurchasing everything. This is more complex than the lift-and-shift model, but it gives you a long-term cost-efficient outcome. CIO Martin Davis says, “you are better redesigning for the cloud if you can. And even then I would focus on building in small manageable chunks focused around business need.”
Many leaders assume the technology that works on-premises is functionally identical in the cloud. This is a mistaken assumption! For instance, I have experienced machine learning libraries that worked on-premises but not for the cloud version of a database system.
In some cases, you might need to keep some data or components on-premises. You might need to comply with a law that requires data be stored in a certain location, or you have legacy databases you want to keep on-premises while testing out your cloud environment.
During a planned migration, replication technology (like Qlik Replicate) can be used to move cloud data into a new database system in real time. If it is a static legacy database, this can be a one-time deal. Individual cloud providers should be able to address your specific questions around legal issues regarding data’s ideal location.
8 Best Practices for Cloud Migration
Once you know what technique fits your business model, you can start the migration process. Here are best practices for on-premises to cloud migration to help you avoid common migration challenges, and keep the process on track and within budget:
1. Create A Budget
Creating a cloud migration budget sounds easier than it really is. When you’re considering the costs, you need to include both:
- Direct and indirect costs from your current on-premises infrastructure
- Estimated costs of using and maintaining that infrastructure in the cloud environment
- Estimated costs from hiring people to convert, transport, and reconfigure data
It’s that last piece that can lead to budget overrun. On average, companies are spending 14% more on cloud migration than anticipated because they lacked a well-defined methodology to deal with the complexity of data migration tasks.
2. Create A Copy Of Existing Data
Before migrating your data, you should create a copy of what you have. As with every other business activity, you need to make sure you protect data’s integrity and availability. With a copy of the data, you have the ability to double-check that all exported data maintains the same scheme and format after the migration. This is beneficial for testing, enabling you to match an identical point in time for where the on-premise data matches the cloud data.
3. Build Out a Data Synchronization Process
Planning your data synchronization process in advance is key to ensuring accurate, secure, compliant data. Basically, it cleans data by checking for errors, duplication, or inconsistencies. Consider data replication first for your data sycnhronization needs. If replication is not available, alternately you can use other methods such as file synchronization.
4. Utilize A Data Catalog To Classify and Label Data
All the applications and workloads you move to the cloud use data. If you have on-premises data silos, then you want to make sure that your data migration doesn’t lead to a budget overrun. Further, you don’t want to move sensitive data to the wrong location which can create a compliance violation and lead to fines.
Using a data catalog that classifies and labels data makes the data migration portion easier because you can make informed decisions and ensure everyone has access to the data they need for their jobs.
5. Design A Transfer Plan For Your Digital Assets
Unfortunately, you’re not going to be able to transfer all data, applications, and workloads at once. Your transfer plan outlines the services you plan to migrate and the consumed resources for each of them. As you build out your transfer plan, you want to consider:
- Transfer priority
- Impact of stoppage time during the transfer
- Critical services that need to remain available
It’s usually a good idea to transfer non-critical services first so that you can test the environment, then add the critical services later.
6. Establish Cloud Migration KPIs (Key Performance Indicators)
Why are you migrating to the cloud? How will you measure and determine success? Every project needs to have metrics that help you understand whether you’re progressing as expected and on budget. Tracking what you’ve budgeted versus what you’re actually paying will help you keep costs in check.
When setting the KPIs for your application migration, you should think about:
- How long it takes pages to load
- Percentage of CPU used
- Conversation rate
- Percentage of memory used
- Disk performance
In terms of your data warehouse migration, you should think about:
- Overall cost in consumption of CPU units
- Storage cost units
- Job execution time
- User query counts
Many organizations migrate to a cloud data warehouse to democratize data, easing access to scale up and accelerate self-service anlaytics. So how do you track this goal? Usery query counts tell you who is querying data and how often. By tracking user query counts, data leaders can track how effective their data democratization efforts truly are.
7. Implement Security Policies and Encrypt Data
Modern organizations handle a lot of sensitive information so you need to make sure that you have the right security in place to protect yourself and your customers. You should set these policies before you migrate your data.
As part of this, you should implement security policies that include:
- Multi-factor and single-sign on authentication (such as Okta)
- Automated password policies
- Access controls
- Separating development and production environments
- Data encryption
- Compliance with any data protection regulations
8. Fit Data Virtualization into the Cloud Migration Strategy
With data virtualization, you put a business-user friendly layer over your data, making it easier for everyone to find, understand, and use. It creates a single location enabling self-service without all the IT complexity, this way everyone can use the business intelligence tools to make informed decisions that drive revenue.
How Alation Supports On-Premises to Cloud Migration
The Alation Data Catalog supports companies through the on-premises to cloud migration process. The migration process is a lot like packing up your house and moving – there’s a lot of planning, sorting, and work. But just like you might bring in people to help you pack your things into labeled boxes, you can use Alation’s Data Catalog to help you sort and plan your data so that you know it goes where you want it to.
Alation’s Data Catalog supports your planning process with machine learning that identifies frequently used data, so you can prioritize what data to move first to the cloud. When you leverage insights from data usage patterns, you ensure relevant data assets are moved and reduce the risk of budget overruns.
You’re also able to create a more efficient strategy that limits business impact. By understanding the effects cloud migration will have on users, processes, and analytical applications, you reduce risk because you know what data needs to be moved first and where data should be migrated. Since Alation users are able to observe and monitor data, people can easily see which assets are being moved, when and where.
Before migration, Alation helps data leaders identify duplicate data so that it isn’t migrated. The platform also supports migration progress tracking. As your data migration gets underway, project leaders can monitor data retention policies, which provide guidance on how much historical data to move.
By centralizing your cloud data migration moving process with Alation Data Catalog, you have a clear and transparent roadmap, as well as a systematic approach for a successful cloud migration.