The public cloud is typically looked on as the great consolidator. Organizations, small and large, use the cloud as a hub for data storage and distribution. Recently, however, there has been a proliferation of edge use cases, and the major cloud providers are not ignoring the trend. Each of the big three (Amazon, Azure, and Google) have released an edge incarnation of their offerings. AWS has Outpost, Azure has Azure Stack Hub and Azure Stack Edge, and Google has Anthos.
The typical edge storage requirement is density and performance but not high capacity. Edge hardware has a limited ability to independently scale-out storage resources. Storage in Azure Stack Hub cannot scale independently from the rest of the hardware resources, so a typical Azure Stack Hub configuration maxes out at under a PB of storage (4 node configuration) or low, single digits PB of storage (16 node configuration). The assumption on the part of the cloud providers is that the edge location will transfer data to the public cloud hub as quickly as possible.
As a result, most edge solutions from the major cloud providers don’t account for a high capacity use case. The need for high capacity can be the result of high network latency, the need for an entirely disconnected or air-gapped architecture, or edge locations that can’t transmit data for regulatory or security reasons. In these use cases, the organization needs an alternative that can scale to multiple petabytes.
Introducing Scality Scale-out Object Storage for Azure Edge
Scality’s RING object storage has been available since 2010, and Scality has been working with the Microsoft Azure team on other projects since 2017. They first came out with an Amazon S3 to Azure Blob translation solution, Zenko Connect, and in 2018 they provided a solution to enable on-premises file stores to connect to Azure Blob Storage with Zenko. Scality has leveraged these capabilities to offer an Azure Blob connector natively within Scality’s RING to an existing file (NFS, SMB) and object (S3) stores.
Scality RING is a software-defined, multi-petabyte, on-premises storage solution that presents native protocols via “connectors.” Scality has connectors for SMB, NFS, FUSE, REST, and S3. In this release, Scality is delivering a Blog API “connector,” which enables customers to place a Scality RING object-store running on the hardware of their choice, next to an Azure edge solution. The same hardware partners that Microsoft has certified to deliver Azure Stack Edge and Hub also approve the Scality solution. This dual certification means the customer can purchase all their hardware from a single source.
Use Cases for High Capacity Edge
While the number of Azure Edge customers that need high capacity is relatively small as a percentage, those that do, need a lot of storage. Energy / Oil and Gas, for example, need to account for Ocean-based platform and remote land-based survey sites that have very high latency issues. These sites collect a lot of data, and transmitting them over a relatively thin and highly latent connection is almost impossible. SaaS providers like email, collaboration, and business app providers may choose to run on Azure Stack and then store data on Scality Ring for better protection and scalability. Defense agencies need mobile data centers to execute tactical deployments, collect and distribute reconnaissance, and surveillance data.
Organizations can also backup Azure stack workloads and data to Scality RING within a single data center. these organizations can leverage Scality RING as a disaster recovery solution. Azure Stack Hub does not yet have advanced DR features, so by backing up Azure Stack Hub in one data center to a RING that is mirrored across two data centers, enterprises can implement a sophisticated Azure Stack HB DR architecture across multiple sites. These organizations can leverage Scality Ring to replicate data between multiple Azure data centers.
There is no doubt that the concept of edge computing is gaining traction. It is interesting, however, to see how aggressively cloud providers are embracing the idea. The challenge is these edge solutions assume that the data footprint is small and temporary. While that assumption is correct for the majority of use cases, there will be use cases where a much larger on-premises storage footprint is needed. There are plenty of S3 solutions, but customers that embrace the Azure Stack portfolio want a native Azure Blob solution. Scality’s integration to Azure Blob as a connector provides a seamless way for Azure Stack Edge customers to store data. It may also have a more general-purpose opportunity for data protection.