Cloud Migration Project Plan: Moving Your Data Center Processes

Most data center managers are locked in a struggle between the need for density and the need to keep equipment cool. Dense hardware tends to generate too much heat, which negatively affects performance. At the same time, some data center cooling solutions have their limits, such as an enormous power consumption bill. For reference, cooling systems can consume 40% of the total power used by the data center.

However, the rear-door heat exchanger (RHDx) offers a solution. With rear-door cooling, it is possible to cool hardware at the level of the individual rack, a capability that provides effective, economical cooling that enables greater density.

Every company is in the cloud, right? Yes and no. While around 94% of enterprises use cloud computing in some way, it’s not as if all corporate IT assets are cloud-based. Roughly 45% of organizations have migrated at least half of their applications to public cloud platforms (according to CDW), yet there are still a lot of assets hosted in on-premises data centers.

The outcomes of cloud migration are not always positive. For those doing multi-cloud, for instance, more than 50% will likely not get the results they expect due to compatibility issues (via Gartner).

Another factor that affects the success of cloud data center migration is a lack of thorough planning. Moving a data center to the cloud is naturally a large-scale, long-term project involving many different people. It is essential to develop a detailed cloud migration project plan that takes many technical and business elements into consideration. The below discusses how to create a good plan and execute it.

What is a Data Center Cloud Migration?

Data center cloud migration is the process of transferring the IT assets running in a legacy data center, including data, applications, and IT workloads to cloud-based platforms like Amazon Web Services (AWS) or Microsoft Azure. The process can also involve migrating data center assets to private-cloud, hybrid-cloud, or multi-cloud architectures.

There is more than one way to execute data center cloud migration. Popular approaches included rehosting, also known as “lift and shift,” which moves workloads to the cloud “as is,” with no changes. Re-platforming is a method of cloud migration that moves most of the same code and application architecture that was used on-premises, but allows for small adjustments to work smoothly on the cloud platform. Refactoring involves more extensive changes to an IT workload as it moves to the cloud, for example, data restructuring or code rewrites.

However you do it, the most important thing to keep in mind about data center cloud migration is that it is as much a business process as it is technical. Business factors such as cost, user experience, and system performance must be top of mind through the planning and execution cycles.

What is a Cloud Migration Project Plan?

A project plan for a cloud migration is a document that guides a cloud migration project through its multiple stages.

It needs to be more than just a document that details how to move IT assets from a data center to a cloud platform. An effective plan is one that takes into account a broad range of strategic, technological, and business considerations. It is the result of an in-depth cloud migration planning process that draws input from numerous stakeholders.

7 Benefits of a Cloud Migration

Why would you want to migrate your data center to the cloud? A move of this kind — if done correctly — offers a number of compelling benefits. These include:

1. Lower costs (and decreased CapEx) —Data centers are not easy to run. They require a lot of dedicated personnel and they consume vast amounts of energy and water, which can turn into an environmental liability. Plus, when we max out the capacity of our very expensive data centers, we face some tricky capital investment decisions.

   A cloud migration means less capital expenditure (CapEx) for hardware and software. It also enables you to employ fewer IT admins to run an on-premises data center. While cloud platforms usually require significant operating expenses (OpEx), they are generally more economical than legacy data centers. It’s important to note that savings aren’t automatic. Workload fit, licensing optimization and data center storage design can all affect the amount of savings that a cloud migration can bring.

2. Ability to scale on demand — Cloud platforms scale up and down quickly, without the need to acquire and stand-up new equipment. This is a major benefit of cloud migration.
3. Agility — Cloud platforms are more flexible than legacy data centers. You can easily shift workloads to new platforms. For example, to bare metal or specialized hardware like ARM-based servers.

4. Security — When you migrate to the cloud, the cloud provider becomes responsible for securing the infrastructure and hardware. This usually lightens your team’s load when it comes to security. It’s essential to understand the “share security model,” because you are still responsible for securing your applications and data.

   With the cloud, you get the benefit of unified management and security interfaces, which can streamline security operations compared to what’s required with on-premises data centers.

5. Disaster Recovery (DR) — The cloud is a great DR resource. DRaaS companies can set up mirrored instances of critical applications, along with geographically distributed backups of data. Confidently know that your data and infrastructure will have High Availability (HA) and real-time failover when critical situations occur. Your business continuity, compliance, and day-to-day operations must maintain performance.
6. Compliance — The major cloud providers have robust compliance policies and processes in place that streamline the process of complying with regulations like GDPR and CCPA.
7. Consolidate data center sites — Companies, especially those that have been through mergers and acquisitions (M&A), might find that they are operating multiple data centers, some of which are underutilized, inefficient to operate, or obsolete. In that case, it’s a good idea to migrate those data centers to the cloud as part of a data center consolidation project plan.

Data Center Migration to Cloud Project Plan

The plan for migrating a data center to the cloud should encompass multiple phases of cloud migration. This is not a one-step, push-button process. Some cloud vendors may present the migration as just a matter of following a few web-based steps to be in the cloud.

Although some of today’s cloud migration tools are extremely powerful, the planning work is still crucially important to understand and develop. The best practice is to plan and execute cloud migration phases. The steps for cloud migration vary, but most plans contain four:

1. Discovery
2. Planning
3. Execution
4. Optimization

Phase 1: Discovery

It‘s a mistake to jump into planning without first undertaking a far-reaching process of discovery. The purpose of the discovery phase is to determine the scope of the data center migration, i.e., what data center elements will migrate, map the migration, and define success.

Discovery also informs the plan by examining business and organization aspects of the data center migration, and many other non-technological elements of the process.

Cross-Functional Conversations

Discovery should involve multiple teams. This is necessary because different groups have their own distinct needs and expectations from the migration. For instance, data managers, infrastructure managers, security teams, disaster recovery teams, and compliance managers will each have their own requirements for a migration.

They each need to be consulted. Also, importantly, the discovery process gives these stakeholders the opportunity to comment on the scope, parameters, and timeline of the migration. Planning a migration the same month as a major compliance audit, for example, would be a mistake. The compliance team should have the right to be involved in this decision.

It’s also a good idea to engage in cross-functional dialogue between organizational units, including those from outside the technology and security areas. For example:

• What does the finance team expect from this process? Do they have metrics of their own that relate to the migration? For example, are they under pressure to cut CapEx or OpEx by a given percentage during the period, and will the migration affect this goal?
• Does the legal department have opinions about the impact of the migration on customer contracts, supplier agreements and software licensing?

Clients may have contractual guarantees of on-premises hosting, among other issues. Liability may be different if data is hosted on a public cloud platform. Software vendors may require different licensing of their products in the cloud than they provide on-premises. For example, can you transfer an Oracle database license to Microsoft Azure, or, if not, what will it cost to use a pre-built Oracle image available on the Microsoft Marketplace?

• Will the migration influence customer experience and user experience? The customer experience team members should be consulted on the impact of the migration.
• Will the migration affect any service level agreements (SLAs), either internally or externally?
• Does the migration create any new recruiting and training needs? Training and development teams may need time to prepare and request budgets.

Assessing your Current Environment

A migration plan builds from a detailed and accurate inventory of data center assets. The discovery phase should include a thorough assessment of the current environment. This will include understanding what hardware is running the targeted workloads, and how that hardware maps to dependent systems such as storage and network.

The inventory will catalogue software, storage, databases, and file volumes. A detailed network map is essential, as are details of operating systems and virtualization.

The inventory serves many different purposes. For one thing, it enables clear project scoping. Are you migrating to 10 virtual machines (VMs) or 100? The inventory also sets the stage for a careful analysis of migration steps and end results. Many problems can emerge in a migration, and it’s great if you can spot them in advance rather than deal with them in the middle of the migration or worse, after the migration.

For instance, there can be small but important differences between operating systems and software versions running in legacy data centers versus those available on the cloud. Seemingly minor compatibility problems can become huge issues. Alternatively, versioning, server capacity, and network topology differences can lead to performance problems. It’s best to see what’s coming, and be prepared.

The data center inventory should be accompanied by a security and compliance inventory. Administrative access controls might be an issue. If cloud administrators require access to systems that were once restricted only to application managers, that could break an important security control. The Privileged Access Management (PAM) solution should take such changes into account. Additionally, it’s a good idea to map out which areas of security are your responsibility under the “shared responsibility model.”

How does this look in practice? A regulated organization might initially plan to migrate a broad set of workloads to the public cloud, but the discovery phase soon reveals application dependencies, licensing constraints, and compliance considerations that makes a phased hybrid approach more practical.

Rationalize your Current Workloads

Another key step in the Discovery Phase is to assess which data sets and workloads are best suited for the cloud environment, and which data sets and workloads you may want to keep on-prem.

There are many reasons to keep certain workloads on premises. Think highly classified financial or government data, personal healthcare data, and integral patented technologies. Not to mention proprietary data that should have extra security and controls to not only meet compliance regulations, but also protect critical organization intellectual property (IP). Critical business workloads that drive financial transactions like retail providers, banking, and brokerage workloads could be better suited for on-premises environments vs. the cloud.

Other applications or lower-security risk workloads may be better suited for the cloud environment. For example, accessing Microsoft’s M365 suite of applications could be a great fit for the cloud.

Identifying Your Goals and Success Metrics

The discovery phase is also the moment to establish your project goals and key performance indicators (KPIs). Goals and KPIs answer two important questions you will likely be asked by senior people: 1) What is the purpose of this (costly, time-consuming, and distracting) project? And, 2) How will you know if you succeeded?

It’s smart to set clear goals and do so in consultation with key stakeholders. Is the migration meant to improve network latency or application response rates? Is the goal to cut costs? If so, what level of cost-cutting will define success? Other typical KPI parameters might include reducing error rates, improving CPU utilization, and cutting downtime.

Phase 2: Planning

Your data center cloud migration plan will likely not be a static document. It’s not as if you will write it, get feedback, lock it, and get busy on execution. Rather, it will be an iterative process that leads to changes as the plan comes together — and continues as the plan is executed.

For some, it’s more of a cloud migration roadmap than a plan. You know where you are heading, but the route may change as new information inevitably surfaces.

Making Practical Decisions

The cloud migration planning process comprises a set of decisions, each of which answers an important, practical question:

• What kind of cloud architecture are you setting up or moving to? This may be predetermined based on earlier choices, but you still must decide if you’re doing a simple move to one platform or migrating to a multi-cloud environment or hybrid cloud, where some assets stay in on-premises infrastructure but link to the cloud. This choice will have a big impact on the details of the migration plan, not to mention the potential difficulties of managing these environments in future, particularly with hybrid cloud management.
• What kind of migration are you doing? e.g., “lift and shift” or re-platforming.
• Do you need to select a cloud provider or work within an existing relationship? It’s likely that you already know which cloud platform will be the destination for your migrated data center, but you will still need to make a clear choice.

From there, you will need to discuss the specifics of your migration with that provider, or a qualified consultant, to determine the optimal server choices, network options, and configurations. The major cloud providers have very sophisticated offerings now, like VMware vSphere on Google Cloud Platform (GCP).

• What’s the timeline you want to stick to? How long will this take? Keep in mind that the execution will likely be in stages, so the whole process could take months.
• What will you do if something goes wrong? This is an under-appreciated aspect of data center cloud migration. There is always the chance that the migration won’t work right, or at all, when you first try it. You need a fallback position, e.g., a failover instance you can bring online if you need to stop the migration, assess the problem, and start over again.

One idea that may be helpful at this stage is to consider keeping major system-specific migration projects separated from a general data center migration. For instance, moving an SAP landscape to the cloud is a huge undertaking with numerous platform-specific challenges. It is also a migration that is best performed with specialized tools leveraged by specialized consultants. It’s optimal not to lump an SAP migration in with other migrations.

Build a Cloud Migration Timeline

As you gain clarity on the nature of your cloud migration and your key objectives, you can form a step-by-step plan with a timeline. One recommended practice is to migrate less critical systems in the earlier stages. This might mean identifying non-production systems that you can migrate to test the process without risking a major business disruption if the migration doesn’t go right.

A time-lined step-by-step data center cloud migration plan might look something like this:

• Weeks 1-6 — Discovery: interview key stakeholders and department leaders to structure goals and migration requirements.
• Weeks 7-10 — Plan development: prioritize lower complexity, security, and risk-level data and workloads for a phased migration approach to the new cloud environment. Plan out sequential phases to migrate more complex data sets and workloads that are suited to the cloud over a manageable timeline.
• Weeks 11-16 — Prepare for Stage 1 migration test: Double-check data and workload dependencies amongst the larger enterprise to make sure that the correct workloads are moved in a logical, sequential order that will not disrupt the business.
• Week 17 — Stage 1 migration test: Simulate, test, and ensure the cloud environment is ready for the migration.
• Weeks 18-24 — Prepare for Stage 2 (Production system)
• Week 25 — Stage 2 migration/cut over
• Weeks 26-30 — Prepare for Stage 3 (Production system)
• Week 31 — Stage 3 migration/cut over
• Weeks 32-36 — Post-migration evaluation
• Weeks 37-40 — Optimization and plans for future improvements

Phase 3: Execution

With the plan in hand, now is the time to execute. Execution usually starts by moving infrastructure components, such as identity and access management (IAM), firewall rules, and networking configuration.

Success depends on remaining agile and responsive. Things are not going to go as planned, which is typical. How you respond to the inevitable challenges will determine how well your execution goes overall.

It’s also wise to test every step of the execution process before cutting over. You will probably want to create a “sandbox” in the cloud where you can migrate a cloned version of the target system and see how it works.

For example, a customer might use a staged migration approach, at first validating a non-production environment in the cloud before moving a business-critical production workload, reducing risk and giving teams confidence in the cutover plan.

Phase 4: Optimize

The migration doesn’t stop with the cutover. The plan should include a period of assessment, followed by optimization. Even if the migration went well, you can probably do better next time. Realistically, there will be performance issues to evaluate and correct.

Data Center Migration to Cloud Checklist

The following cloud migration checklist for the optimization phase could be helpful:

• Check in with all stakeholders on their levels of satisfaction with the migration project and the outcomes
• Measure system performance in the cloud
• Prepare a remediation plan for underperforming systems
• Audit the security of the cloud environment, including penetration testing, and remediate any deficiencies
• Audit the cloud instances for compliance and remediate any deficiencies
• Conduct a retrospective analysis on the overall migration project, covering what went right and what went wrong
• Prepare a list of ideas for subsequent migrations, to improve processes for next time

The optimization stage depends on having the right infrastructure monitoring tools in place — and setting up KPIs to track.

Decommissioning and Technical Debt Retirement

When fully migrated off legacy on-premises hardware, there is still work to be done. Legacy equipment must be shut down properly, securely data-cleansed, and properly recycled.

Data protection and back up (DPaaS) efforts should be audited before any final hardware cleansing is initiated.

When ready, IT Asset Disposition (ITAD) services should be engaged to ensure your data continues to stay secure and is fully cleansed in a controlled environment to fulfill compliance and organizational security requirements.

Once the ITAD efforts are complete, you should ensure that any hardware service support for those disposed devices is cancelled. Software licenses for legacy on-platform devices most likely can be terminated as well. These efforts ensure that you retain budget to support your new cloud environment.

How to Avoid Common Pitfalls when Building a Cloud Migration Roadmap

Errors and unexpected problems are the norm with even a well-planned migration. A growing body of best practices helps you avoid common pitfalls when building a cloud migration roadmap. Here are some of the most important:

• Make the business case for migration — Senior leadership are going to want to know why they’re being asked to spend money on the migration.
• Map dependencies — Applications, databases, storage, and other systemic elements often have hard-to-spot dependencies. Check for them and make sure you have incorporated them into your plan. Specialized tooling can help with this process.
• Analyze costs carefully — Cloud migrations can spring unexpected costs on you. For example, cloud storage often comes with complex and opaque pricing for actions like API calls. It’s wise to get a firm grip on all expected costs in advance of doing the migration. Remember that costs need continuous oversight and not just one-time monitoring.
• Make sure your data is fully integrated — Migration may include the merging of data sets. Invest in tooling and resources to make sure that you won’t have data interoperability issues in the migration.

Choose Park Place for all Phases of Cloud Migration

If you’re ready to migrate your infrastructure to the cloud or to a hybrid set up, Park Place Technologies can be your partner to guide you through every step of the way, with our cloud migration managed services.

We have a proven framework from initial assessment to ongoing technical support once the migration is complete. We ensure that your organization achieves measurable improvements in long-term operational efficiency, and once migrated to the cloud, our private Park Place PowerCloud or Park Place xCloud or public cloud managed services keep your systems running optimally with 24/7 monitoring and management.

Contact us today to learn more about our range of cloud services.