Everyone everyone is talking about cloud as if it had become an obligation. But is it suitable for all information systems? Migrating to cloud is all very well, but why, and what kind of cloud ?
1. The benefits and risks of migrating to the cloud
1.1 Advantages
Here are some scenarios where switching to cloud is beneficial:
- Your application is experiencing an increase in traffic and it becomes difficult to scale resources on the fly to meet the growing demand.
- It is becoming increasingly difficult and expensive to meet your growing storage needs.
- Monitoring and updating the underlying server software is a time-consuming but essential process that requires periodic and sometimes immediate updates.
- Capex to Opex: cloud moves IT spending to a pay-per-use model.
1.2 Risks of migrating to the cloud
Although your specific environment determines the risks that apply to you, there are certain disadvantages associated with cloud migration that you need to consider:
- If your application stores and retrieves highly sensitive data, you may not be able to keep it in cloud. Similarly, compliance requirements might also limit your choices.
- If any of the technology you currently rely on is proprietary, you may not be legally able to deploy it at cloud.
- The design and architecture of your particular application may not completely follow the distributed architectures in cloud.
2. Questions to ask yourself before migrating to the cloud
2.1 Which model to choose? IaaS, PaaS or SaaS ?
Read our full article on IaaS PaaS SaaS
IaaS is the ideal solution for companies that don't mind hosting their applications in third-party data centers, but prefer to outsource the maintenance of their physical infrastructure to focus on development, deployment and monitoring.
However, if you prefer your applications to be portable, you can simply drop your code on a robust PaaS platform that provides a complete infrastructure environment. Adopting a PaaS solution will also reduce your time to market. That's because since the PaaS will come preloaded with most of the software needed to run, all you need to do is deploy the top layer of your application and, in some cases, just the application binaries.
SaaS is a delivery model in which centrally hosted productivity software is licensed on a subscription basis.
2.2 Cloud public, prived or hybrid?
There are three types of accommodation:
- Cloud public: your resources are fully hosted by one or more providers of Cloud, such as AWS, Azure and GCP.
- Cloud private: you create your own private Cloud using a platform like OpenStack.
- Cloud Hybrid: your resources are distributed across private and public platforms. It's the best of both worlds.
2.3 Which providers? AWS, Azure or GCP?
To choose, you should look at the proposed features and the costs.ûcosts. The 3 major providers of cloud offer price simulators to help you compare.
3. Lhe stages of migration to the cloud
3.1 Is your application cloud ready ?
For a successful successful migration, the application must follow a distributed architecture and be scalable by by its design. Tools such as PaaSLane and Cloudamize can help you assess the readiness of your applications for cloud.
Most applications use third party services (web services, external storage, mail servers, authentication services...). It is essential to analyze the impact that your migration to cloud will have on these dependencies.
3.2 Migrating the database
Data is the most important element of an application. Migration can be complex if the data is sensitive or numerous. Vou need to ensure that your migration methods are reliable and have rollback capabilities to deal with any unexpected chaos.
3.3 Making a prototype (POC = proof of concept)
The POC allows to better understand the challenges you may face. Here are a few things to consider during your POC:
- Performance comparison with your existing application.
- Levels of complexity involved in the application migration.
- Network challenges that need to be addressed.
- Reliability.
- Evaluation of the support to cloud.
4. Lhe pitfalls to avoid
4.1 Do not estimate the time and cost of migration
You should estimate the time required to put all your data into the cloud in tenant the ability of your IT team to reconcile the migration with their daily work.
Hosting data on cloud is generally less expensive than using on-premises servers. Your initial cost analysis should take into account the savings from migrating to cloud, as well as the estimated cost of migration.
4.2 Migrate sensitive or critical data first
Before you migrate anything, first plan for the non-critical data you can send to the cloud. Once you have a better understanding of the the cloud architecture usedarchitecture, you can start sending important information.
4.3 Do not maintain security during the migration process
Later data is extremely vulnerable to cyberattacks while in transit. Make sure your data is secure when it is transferred between your servers and cloud.
4.4 Do not train your IT team
Each provider cloud has its own architecture and you have toy architecture. Don't start migrating data to cloud until your team has a good understanding of how your cloud system works.
Everyone is Everyone talks about cloud as if it had become a requirement. But is it suitable for all information systems? Migrating to cloud is fine, but why, and what kind of cloud ?
1. The benefits and risks of migrating to the cloud
1.1 Advantages
Here are some scenarios where switching to cloud is beneficial:
- Your application is experiencing an increase in traffic and it becomes difficult to scale resources on the fly to meet the growing demand.
- It is becoming increasingly difficult and expensive to meet your growing storage needs.
- Monitoring and updating the underlying server software is a time-consuming but essential process that requires periodic and sometimes immediate updates.
- Capex to Opex: cloud moves IT spending to a pay-per-use model.
1.2 Risks of migrating to the cloud
While your specific environment determines the risks that apply to you, there are some disadvantages associated with cloud migrations that should be considered:
- If your application stores and retrieves highly sensitive data, you may not be able to keep it in cloud. Similarly, compliance requirements might also limit your choices.
- If any of the technology you currently rely on is proprietary, you may not be legally able to deploy it at cloud.
- The design and architecture of your particular application may not completely follow the distributed architectures in cloud.
2. Questions to ask yourself before migrating to the cloud
2.1 Which model to choose? IaaS, PaaS or SaaS ?
IaaS is the ideal solution for companies that don't mind hosting their applications in third-party data centers, but prefer to outsource the maintenance of their physical infrastructure to focus on development, deployment and monitoring.
However, if you prefer your applications to be portable, you can simply drop your code on a robust PaaS platform that provides a complete infrastructure environment. Adopting a PaaS solution will also reduce your time to market. That's because since the PaaS will come preloaded with most of the software needed to run, all you need to do is deploy the top layer of your application and, in some cases, just the application binaries.
SaaS is a delivery model in which centrally hosted productivity software is licensed on a subscription basis.
2.2 Cloud public, prived or hybrid?
There are three types of accommodation:
- Cloud public: your resources are fully hosted by one or more providers of Cloud, such as AWS, Azure and GCP.
- Cloud private: you create your own private Cloud using a platform like OpenStack.
- Cloud Hybrid: your resources are distributed across private and public platforms. It's the best of both worlds.
2.3 Which providers? AWS, Azure or GCP?
To choose, you should look at the proposed features and the costs.ûcosts. The 3 major providers of cloud offer price simulators to help you compare.
3. Lhe stages of migration to the cloud
3.1 Is your application cloud ready ?
For a successful successful migration, the application must follow a distributed architecture and be scalable by by its design. Tools such as PaaSLane and Cloudamize can help you assess the readiness of your applications for cloud.
Most applications use third party services (web services, external storage, mail servers, authentication services...). It is essential to analyze the impact that your migration to cloud will have on these dependencies.
3.2 Migrating the database
Data is the most important element of an application. Migration can be complex if the data is sensitive or numerous. Vou need to ensure that your migration methods are reliable and have rollback capabilities to deal with any unexpected chaos.
3.3 Making a prototype (POC = proof of concept)
The POC allows to better understand the challenges you may face. Here are a few things to consider during your POC:
- Performance comparison with your existing application.
- Levels of complexity involved in the application migration.
- Network challenges that need to be addressed.
- Reliability.
- Evaluation of the support to cloud.
4. Lhe pitfalls to avoid
4.1 Do not estimate the time and cost of migration
You should estimate the time required to put all your data into the cloud in tenant the ability of your IT team to reconcile the migration with their daily work.
Hosting data on cloud is generally less expensive than using on-premises servers. Your initial cost analysis should take into account the savings from migrating to cloud, as well as the estimated cost of migration.
4.2 Migrate sensitive or critical data first
Before you migrate anything, first plan for the non-critical data you can send to the cloud. Once you have a better understanding of the the cloud architecture usedarchitecture, you can start sending important information.
4.3 Do not maintain security during the migration process
Later data is extremely vulnerable to cyberattacks while in transit. Make sure your data is secure when it is transferred between your servers and cloud.
4.4 Do not train your IT team
Each provider cloud has its own architecture and you have toy architecture. Don't start migrating data to cloud until your team has a good understanding of how your cloud system works.
