WANdisco is known in the Big Data world for its Fusion product, that keeps multiple Hadoop clusters in sync. Now the company is leveraging that technology to enable multi-cloud data solutions, with the aptly named LiveData for MultiCloud. And the general availability of that product is being announced today.
While many in the industry agree that multi-cloud is a major trend, most see Docker containers and Kubernetes orchestration as the enabler. WANdisco has a different opinion. The company sees high-volume migrations and sharing of data in real-time, across clouds, as the true multi-cloud enabler.
Sure, multi-cloud aficionados can take a lowest common denominator Infrastructure as a Service (IaaS) approach and move containers around interchangeable Kubernetes services (like Amazon EKS, Microsoft AKS, and Google GKE). But WANdisco says its approach allows customers to use Platform as a Service (PaaS) products on different clouds and still let them share data.
Mix and match
For example, an Enterprise customer may wish to use Amazon's DynamoDB for NoSQL data management, but may find Microsoft's new Azure Data Explorer (formerly known by its project code name, "Kusto") to be attractive for time series analysis on the same data. Using a patented, high-performance blockchain-driven technology that WANdisco calls DConE, LiveData for MultiCloud facilitates this. It does so by managing the real-time, active-active replication of data between cloud object stores, for example between Amazon S3 and Azure Blob Storage or Azure Data Lake Storage. The product also supports Google Cloud Storage and facilitates on-premises/cloud hybrid topologies as well.
While technologies like change data capture (CDC) can work for unidirectional data replication, it's bi-directional (or multi-directional), continuous replication that's necessary for multiple applications to read from, and write to, the same data. WANdisco says it provides for just that, at petabyte (PB) scale. Which is good, as WANdisco's CEO, David Richards, told me the company has one customer that now produces 4 PB of data per day.
PaaS without the lock-in?
My ZDNet colleague, Jason Perlow, made the point recently that no matter how customers may try to avoid cloud lock-in by using Docker and Kubernetes, sooner or later, they're going to succumb to using the public cloud providers' PasS products. Though these services may be proprietary, they are also often where the innovation lies and, as such, permit customers to reduce the cost of development.
Also read: Game of Clouds: Lock-In is Coming
WANdisco would say its solution lets the cloud providers' customers avail themselves of PaaS products and avoid lock-in. Richards explained to me that LiveData for MultiCloud essentially implements a globally consistent file system across cloud object stores, and that's what makes all this work.
While this approach could potentially double (or even triple) customers' storage costs, Richards feels the cloud providers price their storage services so low, that they act as loss-leaders to encourage stickiness for the other services. He says that multiplying those costs is reasonable, if it means the stickiness is mitigated. It's a reasonable argument. Now let's see what the market says.