As we move to a hybrid cloud world, data mobility will become increasingly important. I’ve been talking about this issue for some time, as a key feature of a hybrid cloud strategy. Without the ability to easily move data between cloud providers and/or geographical locations, all we end up with are isolated islands of public cloud applications.
- Modern Storage Architectures: Datrium
- Technology Choices for Data Mobility in Hybrid Cloud
- Hybrid Cloud and Data Mobility
With the introduction of Automatrix, Datrium brings together five important components that create a data mobility strategy that just might deliver on the promise of hybrid cloud.
Application Mobility
Today we have application mobility. We can spin up and create a virtual machine, virtual instance or container, pretty much instantaneously, in pretty much any cloud. Using the right tools, applications as code is relatively easy to deliver.
I say relatively because work is required to build out the application definitions, implement security and otherwise network the components together. The challenge though is putting data in the right place to ensure applications can use it.
Data Mobility
Why is data mobility such an issue? We know that data has inertia (some people call it gravity, but I think inertia is more accurate). It takes time and effort to move data around. When data is in flight, we have a consistency issue; either we can’t use it or have to use methods to manage updates and keep data coherent. Public cloud providers also charge for data egress, putting a dollar value on moving applications around.
Moving applications between clouds isn’t trivial – so is it worthwhile?
Disaster Recovery
Moving applications between disparate public and private clouds should be done for a good reason. Disaster Recovery is cited as one example of how the Automatrix architecture enables data and application mobility. Today, the Automatrix DR process allows VMware virtualised applications running under Datrium DVX to move to another on-premises deployment. DVX provides the underlying framework to store, de-duplicate, encrypt, replicate and automate the DR process.
The last point is, in many ways, the final piece of the puzzle in delivering full mobility and one of the five components of the Automatrix architecture.
Five Tenets of Mobility
Datrium has identified five components for data and application mobility.
- Primary storage – we can take this as a given. Applications, even container-based ones, need to be stored somewhere.
- Backup – data needs to be replicated in order to provide resiliency for recovery.
- Encryption – specifically, consistent encryption at the data layer, so savings can still be made in data services and in moving data around the network
- Mobility – the ability to move data across a wide-area network in as an efficient way as possible.
- Automation – or in the case of Datrium, the automation of the entire DR process.
Some of the components to achieving mobility are already in the Datrium platform. DVX provides primary storage, encryption, de-duplication and mobility. Cloud DVX provides backup and a source of data for DR (backup into AWS S3). ControlShift (formerly Project CloudShift) delivers the automation and validation of workload migration. This includes automated compliance checking and test failovers.
Use Cases
Disaster recovery is being presented as an initial use case for Automatrix. From a cost and operational efficiency perspective, this makes complete sense, especially when combined with public cloud solutions like VMware Cloud on AWS.
Keeping an entire set of replica hardware, just for the rare occasions that failures occur, is an expensive process. Instead it makes more sense to be able to pay for DR when it is needed. This what Automatrix with VMware Cloud on AWS offers. Data is replicated to AWS S3 with Cloud DVX then re-hydrated back into an AWS-based VMC if a disaster occurs.
The saving here is clearly in not having to pay for a second set of hardware. After all, insurance shouldn’t be as expensive as the insured product itself.
Mobility Benefits
I see a much wider application for the technology that Automatrix can offer. Imagine using the solution to perform data centre consolidation. It could also be used in data centre refresh, instead of purchasing swing kit. In both scenarios, applications could be moved into the public cloud as well as to other on-premises locations.
In the future, if/when support for VMware on Azure is available, Automatrix could remove the need to run any on-premises infrastructure at all. This is achieved, crucially, without having to transform any applications into being cloud-native.
Futures
All of the above discussions assume that the application abstraction layer is delivered by VMware vSphere and ESXi as the hypervisor. As applications migrate to containerisation, how does Automatrix fit here? DVX already supports Docker and will support Kubernetes in the future. As VMware has become the de-facto VM virtualisation layer, so Kubernetes is cementing itself as the standard for container orchestration. Automatrix will support Kubernetes in a future release.
Cost Savings
IT organisations that are currently heavily invested in VMware-based server virtualisation can, without doubt, benefit in using Automatrix. Check out our recent sponsored podcast in which Tim Page highlights customers that chose DVX and Automatrix, simply for the storage consolidation savings it offered.
For further background, we spoke to Datrium last year about DVX and CloudShift (now ControlShift), which are available as Storage Unpacked podcasts.
The Architect’s View
If Datrium can offer the same levels of flexibility in data movement for containers that are being achieved for virtual machines, then Automatrix could represent an all-encompassing solution for data mobility. This is especially true if data within VMs can be easily be detached from the VM itself and presented to a containerised version of the same application.
What’s not clear is where Datrium heads next. Is there value in supporting non-VMware platforms like OpenShift? How would Automatrix integrate into supporting cloud-native virtual instances, if at all?
Perhaps here the opportunity is to present application data to virtual instances, cloud-based Kubernetes or even serverless functions. This means making application data appear local, wherever the application runs. At this point, we might have actually reached true data mobility.
To see components of Automatrix in action, check out this video from a recent Tech Field Day event (and other videos here).
Disclaimer: Datrium is a client of Brookend Ltd.
No reproduction without permission. Post #2CAC.