As discussed in earlier sections, originally, our focus was mainly on greenfield applications and how to leverage cloud-native capabilities like serverless, containers, microservices architectures, CI/CD patterns, and so on. However, in typical enterprise environments, most customers already have significant investment in their existing on-premise or colocation environments, so those workloads also need to be moved to the cloud to benefit holistically. To enable the same, Google Cloud offers some native services as well as partner offerings which can be leveraged across various stages of migration. Broadly speaking, Google suggests four different phases in any migration project, which includes assessment, planning, network configuration, and replication.
For most the part, during the assessment and network configuration phases, the onus is on the customer to look at the appropriate tooling, developing automation that can enable the discovery of existing workloads in on-premise environments as well as the creation of matching the network config on cloud, which can enable a seamless migration path.
For the planning phase, Google has a set of recommended partners, such as Cloudamize, CloudPhysics, and ATADATA, which can be leveraged to inspect the current on-premises environment and accordingly map them to the correct cloud services as well as right-sized instance types for best performance.
Similarly, for the actual VM migration phase, Google has a set of recommended partners, like CloudEndure, Velostrata, and ATADATA, which can again help replicate on-premise virtual machines directly to the cloud. Apart from these partner offerings, Google Cloud also offers an option to directly import the virtual disks using its native service capability, however that's not an optimal option for large scale migrations and partner products offer better capabilities there. You can find more details on the VM Migration aspects at the following link: https://cloud.google.com/compute/docs/vm-migration/.
Other than VM migrations, another very important aspect to consider in large scale migrations is around data migration. To do that, there are a few different options, as follows:
- Cloud Storage Transfer Service: This service helps you transfer data from other clouds (like AWS) to Google Cloud storage buckets using HTTP/HTTPS-based interfaces. You could do this directly using the Google Cloud Console, REST APIs, or even using Google Cloud Client API libraries. More details can be found at https://cloud.google.com/storage/transfer/.
- Google Transfer Appliance: Like the AWS Snowball service, Google Cloud also offers a physical appliance which you can lease from Google to transfer large amounts of data to the cloud. At present, it's available in 2 different sizes – 100 TB and 480 TB – which you can request for a few days to connect to your local storage systems, transfer data, and then ship back to Google to transfer it in one of your cloud storage buckets. Before being stored on the Transfer Appliance, all captured data is deduplicated, compressed, and encrypted with an industry standard AES 256 algorithm using a password and passphrase that you specify. With this service, you save on both cost and time, as this helps you quickly migrate the data without using your internet bandwidth. More details can be found at https://cloud.google.com/transfer-appliance/docs/introduction.
- Google BigQuery Data Transfer Service: In most cases, customers are also using many SaaS applications, like Adwords, DoubleClick Campaign Manager, DoubleClick for Publishers, and YouTube. In order to analyze the data from these services, customers can use the BigQuery Data Transfer Service to directly transfer their data to BigQuery and build a data warehouse to run analytical queries. This service ensures continuous data replication at scale, and to visualize the trends, customers can use ISV offerings like Tableau, Looker, and ZoomData, on top of the BigQuery instance. More details can be found at https://cloud.google.com/bigquery/transfer/.