CLOUD MIGRATION & KEY CONSIDERATIONS
Every organization is moving towards digitalization in some form or other and one of the commonly used step is Cloud Migration to achieve this. Most of cloud providers offer various tools & services which accelerate digitalization process and enable customer to take decisions quickly based on data available in diverse IT systems.
There are multiple factors which customer and cloud migration team need to consider while planning for cloud migration. This article details some of the key considerations during planning phase of cloud migration projects. Some of these may not be applicable for simple Lift & Shift (or Rehosting) migration approach, but knowing these will help you to take informed decision in early phase.
This article doesn’t require any prior knowledge of cloud technology. Even though I have used technical terms from Amazon Web Services (AWS) in some places to provide further insight, the concept remains the same irrespective of cloud provider such as GCP, Azure etc.
Let us start with the most important considerations:
I. Network Bandwidth & Data Volume: This is one of the most important criteria before planning any cloud migration and can impact most of decisions during planning phase. There are alternative ways to migrate data from on-premise to cloud based on volume of data.
We can understand this factor using below sample scenario:
a) Need to transfer 10 Petabytes of data within 1 month
b) Need to transfer 10 Terabytes of data within 1 month
c) Need to transfer 10 Gigabyte of data within 1 month
Needless to say that each use case will require different data transfer technology to meet specific scenario.
Most of cloud providers have different options to address these situations. If you are working with AWS, you may choose Direct Connect (dedicated network between On-prem & AWS for steady data transfer), Snowball (used for one time Petabyte scale data transfer), Site2Site VPN (IPSec protocol, over internet) or upload your data directly to S3 (using multi-part upload over internet).
You may identify the appropriate data transfer technology as per your specific need during planning phase of cloud migration. This may also help to access application in cloud post migration.
II. Business Criticality & Outage duration: Every cloud migration will have some disruption to the business and the outage (or downtime) should be discussed early in planning phase. Some business may agree for an outage in weekends, but some may have high transactions in weekend and agree for outage during weekday/night.
Let’s see below scenario to understand this better:
a) Outage can be allowed for maximum 10 min for a critical application.
b) Outage can be allowed for entire weekend for an internal application.
The maximum allowed outage duration will impact the strategy and cost of cloud migration. Normally if the outage is allowed only for few minutes, there will be higher migration cost due to parallel systems running between On-premise & Cloud environment before actual cutover.
III. Use of Licensed Software There can be different migration approach depending upon specific licensed software used at on-premise and how much flexibility does your licensing contract offer with respect to its usage in cloud environment.
Let’s see below alternatives to understand it:
a) No change in license and license can be re-used on cloud (ex: Oracle to Oracle — for specific condition)
b) Moving from individual license to cloud managed licensed product (ex: Oracle to AWS RDS for Oracle)
c) Moving from licensed software to open source software (ex: Oracle to Postgres)
d) Moving from licensed software to SaaS product with same provider (SAP HCM to SAP Success Factor)
e) Purchase of new license as on-prem license cannot be used in cloud.
This is obvious that each use case will need a separate strategy for migration and impact cost accordingly.
Many customers choose to move away from licensed software to open source software (if possible) during cloud migration. This is useful in reducing the future cost and dependency on a software provider after migration.
It is strongly advised to go through licensing contract documents and discuss with your legal team (about small prints) & license provider to avoid any risk related to non-compliance, legal disputes or claims.
IV. Variability in Application’s Load The selection of cloud native tools & services will also depend on the variability in load of application.
Let’s define variability of application load for a particular time-period as,
Variability = (Max application load)/ (Min application load)
Application load can be measured in term of no. of incoming traffic, no. of concurrent users, no. of API calls (or anything which you measure for your application load based on your specific business) for a defined time period say hourly, daily etc.
Now, suppose the application load are as below:
a) Variability is in range of ~10,000
b) Variability is in range of ~ 100
c) Variability is in range of ~1
The target Cloud architecture will depend on the variability in application load and specific services should be used to achieve better results for different scenario.
Typically if the system/application is used with high variability, you may use serverless architecture which manages scalability without any prior planning.
This criteria will impact the design of cloud architecture and hence should be considered during cloud migration planning phase.
V. Adherence to Cloud Native Architecture: One of the most important reason to migrate to Cloud is to use cloud native architecture. There are multiple technologies which should be considered during planning for cloud migration such as:
a) Elasticity to improve cost
b) Auto-scaling to handle variability in application load.
c) Infrastructure as Code (IaC) to manage infrastructure resources using codes.
d) Managed Services to reduce the IT labor cost.
e) Serverless services to manage scalability, elasticity without any manual intervention.
f) High Availability to increase reliability
g) PaaS model to increase speed to market.
h) Decoupling to manage growth of individual components.
i) Caching to reduce the latency and cost.
j) Microservices to manage change and complexity
and so on..
These factors may also impact the migration strategy but help to realize the benefits post migration.
VI. Prioritization & Trade-offs:
It is obvious that IT system/application cannot be revamped to realize complete benefits of cloud native technology during migration due to limited timeline, resources, budget etc.
There may be specific needs of business and hence some features of cloud may have higher priority than others. Customer may review the target application architecture during cloud migration and should include relevant cloud technology to ensure high availability, elasticity, cost reduction etc.
There is always a trade-off due to competing business needs and prioritization of requirements. For example, there will be increase in cost if you plan for high reliability to avoid a single point of failure.
But, there are some mandatory requirements such as Security & Compliance, Data Protection, Legal Regulations, Compatibility of OS/Software/Application versions etc which should be considered for target cloud environment during planning phase.
VII. Cost Consideration:
This is one of the most important factors and cloud migration may depend upon the IT budget available for migration (and expectation of cost benefits post migration). I have shared the cost improvement related details in my earlier article at:
https://medium.com/@b95rkumar/migration-to-cloud-cost-reduction-a3eac59254c1
Overall, these are some of most important considerations during the planning phase of Cloud migration. Customer as well as cloud migration team should discuss these factors and include these in their migration strategy as per their need, budget, schedule etc.
I am sure that if you discuss these considerations early in your planning phase, the cloud migration journey for your organization will definitely be smooth and benefits of cloud migration can be realized quickly.
