Object storage stands to unlock cost-effective archive and disaster recovery, but most object storage systems can’t keep up with the pace that backup applications can send data to them. Another pain point is that some mission-critical applications may need to be rewritten to support object storage, which is expensive and risky.
Tributary Systems
Tributary Systems focuses squarely on facilitating data availability for mission-critical enterprise compute platforms, in the form of integrated solutions for backup, replication and disaster recovery. Its portfolio includes customized tape automation solutions for HPE NonStop, IBM Mainframe, Power and Pure Systems, as well as all open environments such as Linux. It also offers its Storage Director backup virtualization product, which provides backup and disaster recovery for IBM’s Cloud Object Storage (formerly Cleversafe) and on a more limited basis, Hitachi Content Platform, object storage platforms, as well as public cloud object storage services from Amazon, Azure, Google, and IBM.
Tributary Systems Storage Director and Object Storage Support
According to Tributary Systems, its Storage Director platform can virtualize data from any host storage system, operating system or backup application. Storage Director can then back up that data to, or recover it from, any storage device or S3-compatible object store. Multiple targets can be used as backup and disaster recovery targets, on a policy-driven basis. Among the most unique aspects of Storage Director is its ability to support open and proprietary environments (including Linux, Windows, VMware, UNIX, IBM z/OS, IBM iSeries, HPE NonStop) together on a single node.
Storage Director enables data to be archived, post-backup, to the lower-cost object storage target. It was designed to do so with greater speed and efficiently when compared to a direct archive, while enabling customer to continue to follow their pre-existing backup processes. According to Tributary, it can write approximately 11-12 TB per hour per node to IBM Cloud Object Storage, and it can read approximately 8-9 TB per hour per node from IBM Cloud Object Storage. Tributary claims that backup applications provide performance that is only approximately one-third or one-fourth of this.
The Storage Director platform is pre-integrated with a SuperMicro compute node that is provisioned with anywhere from 10 to 600 terabytes (TB) of storage capacity, and that is connected to the object store via 10 Gigabit Ethernet (GbE) networking. It also can connect to any S3-compatible on-premises or cloud object store. The on-premises Storage Director appliance stores metadata and can also cache up to six months’ worth of files. For faster transfers to the object store, it automatically dramatically increases block sizes before they are migrated. This approach accelerates backup and disaster recovery read and write performance. It can also help to reduce the overall cost structure of the storage environment, as many primary storage systems including IBM iSeries cannot integrate with object storage. Furthermore, caching data can reduce egress fees from public cloud service providers, and it can help to avoid the impact of potential service interruptions.
StorageSwiss Take
Object storage has grown far beyond cold, static archive use cases. It can play an integral role in an active backup and restore implementation, supporting crucial data availability, compliance and retention requirements. At the same time, not all applications can simply be lifted and shifted to object storage, and the architecture must be implemented in a way that preserves performance. Tributary Systems offers a path to backup-as-a-service for these applications, as well as a means to access low-cost and searchable object storage without incurring typical latency penalties. Enterprises looking to modernize legacy mission-critical applications that are also integrating more modern, open-source infrastructures in particular will find value in the Tributary Systems solution.