Nobody really enjoys data migration, which is why interest in automation is so high when it comes to highly dynamic virtual and cloud environments. However, automation is only as good as the processes, and the assumptions they’re based on, that are in place. In a conversation with IT Business Edge’s Arthur Cole, RiverMeadow CEO Mark Shirman says enough is known about the cloud to implement full end-to-end automation for the workloads that need it. For the rest, there is a semi-automated tier that reserves key elements of the process for human hands.
Cole: Amid all the talk of the wonders of the cloud, few people mention the migration. What are some of the main challenges enterprises face when moving workloads and applications to the cloud?
Shirman: Most organizations, and even service providers, have been focused on new or ‘greenfield’ servers and applications. Additionally, people are using the cloud for test and development environments. However, this represents the smallest opportunity to migrate to the cloud. Moving existing or ‘brownfield’ servers to the cloud is very difficult and manually intensive, and this friction has slowed down the adoption of the cloud.
At RiverMeadow, we’ve decided to focus solely on this aspect of cloud adoption: reducing the cost and complexity around moving existing servers and giving the enterprises and service providers the freedom to rapidly expand their markets. This moves the entire concept of cloud adoption from a science experiment to a business reality.
Cole: RiverMeadow's SaaS migration platform aims for full automation of the entire process. But since the cloud has such a diversity of platforms, won't this obscure some of the unique offerings out there?
Shirman: We aren’t making a bet on which cloud architecture will win. We just want to give people the freedom to optimize their servers in their chosen environment. We leverage the necessary APIs in the cloud environments within our toolset, so that we can essentially become a Swiss Army Knife for hypervisors and cloud architectures.
In essence, we take the difficult part of the migration and automate it. There will always be a percentage of workloads that fall in the category of full automation. Then there is a set that is semi-automated – situations where the server requires some sort of involvement in the changing of things. We provide a platform that accommodates the full spectrum of migration.
Cole: Is there a danger that end-to-end automation of the migration process will inhibit interoperability of the data environment? How can IT be assured that the newest cloud environment coming online will be compatible with existing services?
Shirman: RiverMeadow works with the standard APIs and interfaces that are available from an assortment of vendors. Compatibility is not an issue. We actually enhance interoperability and portability of workloads between clouds and hypervisors.
The migration process has three components: collect, convert and deploy. Our system collects the image of the server remotely, moves it into a sandbox environment, emulates the environment that it is moving into, and then converts it to the target hypervisor for test and verification. In this way, once it is auto-deployed into the cloud, it is using a fully compatible working virtual machine for that environment.