How to plan for Cloud Migration?

This article is about cloud migration planning. There are three major activities in cloud migration viz. Virtual machine migration, Application migration and Data migration.

Why Cloud?

Cloud is more cost efficient, reliable, easy to use and fast to start. You have to pay only for what you use without any upfront commitment.

Typical Enterprise Challenges:

• Security concerns
• Ineffective asset management
• Wide variety of small & large business applications
• COTS and legacy applications not cloud ready
• Outdated source code
• Hard to baseline application performance
• Understanding new IT Operating model and tools

Enterprise Constraints when Migrating to Cloud:

First goal is to leverage existing investments wherever possible instead of making new because we’ve sometimes uncertainty into cloud solution. Knowledge and training is required to educate the team about cloud technologies. It is good to have prior migration experience. It is difficult on security side if data Classification not fully understood. Understanding of data is more important as it is most critical part of product. It is also considered as a big security loophole if you’ve inappropriate controls over your product data.

Migration methodology:

• Plan: It is mostly discovering applications in current data center and designing its equivalent architecture on cloud. Data center migration assessment is also one of its steps.
• Build: It has transform and transition. Re-hosting or what we call lift and shift. Re-platforming or what we call lift and reshape.
• Run: In this phase you’ll operate and optimize migrated applications along with portfolio optimization.

Minimum infrastructure information:

Number of servers/VMs including RAM, CPU, OS, and boot drive size. Storage mapping to transnational, backup, archival, and log/file system/applications. Geographical location of data center where processing is happening. Data transfer quantity and speed out for networking. Public and Private components with its networking including security requirements.

Backup requirements for each workload that can not be supported by standard cloud storage. You need to identify high availability (HA) requirement wherever needed. If possible, which applications are scalable and which are not. Need to set up disaster recovery (DR) for almost all workloads. In most of the outage the storage IOPS is the bottleneck. So you need to know required IOPS for each workload.

If you’ve separate management and monitoring tool it is good to identify compute requirements of it.

If your application is HIPPA /PCI compliance then it is important to design cloud architect in a way that the compliance rules are not violated at any given point of time.

Sometimes high performance computing (HPC) is not exactly same on cloud as we’ve on-premises because on cloud we don’t have extremely high end machines which again varies provider to provider.

One more thing:

Cloud provider doesn't have its own advanced firewalls like Cisco, Checkpoint etc. Cloud often comes with simple firewalls like security groups and network ACL. So you need to install third party advanced firewalls if required.

Invest in POC early:

Proof of concept (POC) will answer tons of questions along with its implementation dos and don’ts which is very useful to get your feet wet with cloud technologies. It is helpful for you to identify gaps and touch points. POC also gives you the good estimation of migration and runtime cost.

Migration Design & Planning Process:

• Solution design: Conduct workshops to get idea about cloud and its migration strategies. Design well defined target cloud architect once you’ve got enough knowledge of current data center architect. Also get list of tools which are helpful in migration process; of course after POC.
• Migration Plan: First define migration sprint planning and milestones along with migration complexity and efforts. Establish performance and data validation and acceptance criteria. Most important to build migration checklist and execution plan.
• Refine Process: it is recommended to conduct migration pilot for initial trial run. Update tool and process on the basis of pilot outcomes. At last validate assumptions and migration checklist.

Migration Execution Approach:

According to migration pilot implement migration automation tools. It is important to leverage high speed data transfer from data center to cloud for minimum application outage. This can be achieved by extending IT operating model to the cloud. Don’t forget performance bench marking and validation testing.

Migration Execution Process:

• Create Cloud Environment: Prepare future target and deploy core infrastructure services. Setup central control source for accounts, credentials, policy and permissions.
• Prepare On-premises: Prepare on-premise Infrastructure readiness. Capture all relevant applications/images/snapshots as per prioritization list.
• Deploy into Cloud: Deploy captured application/images/snapshots in the target environment with right sizing of cloud resources.
• Migrate Data: Determine data migration approach first. You can also execute parallel run if needed, don’t forget to test data consistency at the end.

Some Migration Tools:

Below are some tools which are helpful in overall migration process.

Please note I don’t have any connection or advertisement agreements with them, they are harmonize with migration process.

• Host Cloning: Racemi, CliQr, DoubleTake, ATADATA etc.
• DR Replication: CloudVelox, CloudEndure etc.
• VM Conversion: Rivermeadow, Zerto, Ravello etc.
• App Container: AppZero, C3DNA, CliQr, UShareSoft etc.

Thanks for reading.