Pure gets into mainframe backups with Model9

Pure Storage’s FlashBlade has gained the ability to store mainframe backups – and make the data available for use in AI/ML and analytics applications – thanks to a partnership with Model9.

Pure’s FlashBlade is a unified file and object all-flash storage array using QLC (4bits/cell) NAND with up to 1.9PB of raw capacity per chassis. Model9’s Shield product takes mainframe backup data, converts it to S3 object storage format, and then moves it to a target object storage device either on-premises or in the public cloud. With this partnership, FlashBlade can be the on-premises device.

Chad Monteith – principal architect, field solutions & strategy, office of the CTO with Pure Storage – said: “Historically, mainframes have been isolated islands of data with limited connectivity to the services and applications in modern datacenters. The joint Model9 and Pure Storage solution modernizes the mainframe with connectivity to open systems platforms, enabling modern, cloud-like economics and application workflows, and bringing a more effective data consumption model to enterprise customers.”

Once the data is in the Pure system it is virtually air-gapped from the mainframe and, like all Model9 target systems, can be used for DR, ransomware recovery and archiving as well as backup.

When the data is inside FlashBlade, a second Model9 product – called Gravity – can run Extract-Transform-and-Load (ETL) procedures against it to generate a subset for use in AI and machine learning applications running on non-mainframes – open systems servers in other words – again either on-premises or in the public cloud. Big data-type analytics apps are another processing target for Model9’s ETL procedures.

In theory the mainframe-originated data could be munched by AI/ML and analytics apps on the mainframe directly – if they have been ported to run there and if the customer can withstand the cost of those mainframe MIPs.

Model9’s (and Pure’s) argument is that it is more efficient – meaning fewer mainframe cycles are needed – to go through this somewhat complicated backup-convert-move-to-Pure-ETL-move-to-open-systems procedure to get the data into the hands of the AI/ML and analytics apps than to do it on the mainframe itself. The open systems servers will generally be owned or operated by the mainframe customers as well, so the Model9-Pure FlashBlade process is, in effect, a gigantic ETL framework scheme to get mainframe data into these servers.

Model9 CEO Gil Peleg explained: “With more and more mainframe customers embracing [hybrid] cloud, and 80 percent of the world’s business data still held in mainframes, it is imperative that companies find a way to integrate their most valuable, mission-critical data with AI/ML and analytics applications while making sure the data is well protected. The Pure Storage and Model9 joint solution allows customers to create a hybrid environment that will enable them to enjoy the best of both worlds, while ensuring business continuity.”

Who knows how this could develop, once Pure has FlashBlade arrays sitting next to mainframes in enterprise datacenters? Talk about having a foot in the glass house door.

Model9 and Pure Storage will co-sell the joint system though systems integrators and resellers in the US and globally.