SwiftStack adds some significant capabilities around public cloud archiving, synchronization, and multi-region clusters – and indicates that much bigger news will be coming in a month.
San Francisco-based SwiftStack has announced its fall product update, their 5.9 release, which focuses on new multi-cloud data management. The enhancements include multi-region erasure coding and improvements to their Cloud Sync tool to allow syncing between different public clouds.
The multi-cloud data management enhancements are more than the typical improvements to capabilities and feeds in speeds that are a regular part of product updates. SwiftStack, since their inception in 2011 to commercialize OpenStack Swift, has been classified as an object storage vendor. Their focus has always been providing the ability to stand up a service like Amazon S3 inside an enterprise data centre instead of the cloud, storing the data on servers rather than in traditional storage. Now, however, SwiftStack is increasingly defining itself as a multi-cloud data management provider, and the new improvements are geared around that aspect.
“We are an object storage solution, but the product really enables multi-cloud data management,” said Mario Blandini, Chief Evangelist at SwiftStack.
Part of the rationale here is also an assessment that object storage hasn’t met market expectations.
“It hasn’t delivered on its promises of performance,” Blandini said “The apps that write to it don’t really exist. Channel partners tend not to sell it. On the other hand, connecting the Google and Amazon clouds efficiently – we are a solution for that. We have a stronger value proposition outside of the storage classification.”
Blandini said that SwiftStack’s status as a born-in-the-cloud vendor always differentiated them from other object storage vendors.
“While we were classified as object storage, we were created in the cloud world,” he said. “Others in our space are pre-cloud and have adapted their technology to the cloud. Data lives longer in the cloud. DreamWorks is a customer of ours, and it typically takes them five years to make a movie. That means that their data typically has to undergo a forklift upgrade at some point. Customers also want choice in clouds. Lots of customers want to buy servers from more than one vendor. It’s the same with the cloud.”
Blandini indicated that the enhancements in this release are significant. They simplify cloud-to-cloud integration and add policy-based rules for multi-regional replication and erasure coding. Yet at the same time, he indicated that much bigger news will be coming a month from now.
“SwiftStack has done a lot between March and now, a lot of cool stuff that we haven’t communicated, but we will be making major news in four weeks, entering a new market,” he said. “This is housekeeping.”
The release improves compute and archiving workflows around the public cloud, with auto-tiering based on IT-defined policies. This lets users consume storage from a single namespace regardless of what cloud it is in, so they can choose which policy is best for their application. SwiftStack now manages the process of repatriating a workload from public cloud to private cloud seamlessly even while applications are online, and maintains uninterrupted access to data.
“These enhancements to policy-based data management make it easy to set up policy automation,” Blandini said. The data structure appears the same to the user, regardless of location, and it behaves like it was still on-prem regardless of location. A simple icon indicates that the data is at rest somewhere else.
SwiftStack’s Cloud Sync, a tool for public cloud archiving which facilitates replication of data from a SwiftStack cluster into a public cloud bucket, has also been broadened.
“You can now sync between two clusters in different regions,” Blandini said. “You could always go from a region to AWS, but not this allows for better changing of policies on the fly.”
Multi-region erasure coding protection has been added for the first time.
“We’ve always had the capability to do multi region clusters – but with replicas,” Blandini said.” Replicas are the equivalent of Raid 1, while erasure coding is Raid 5. Now instead of being limited to a single geographic region, multi-region erasure coding capability lets you achieve efficient capacity ratios, and lets you more easily get access to data even if a whole node goes offline.”
Cisco is a major strategic partner of SwiftStack – because SwiftStack’s technology stores data on servers, not on storage.
“Cisco is our number one reseller,” Blandini said. “We are having great success with them and their partners.”