It’s no longer about data storage; it’s all about data management. Today, if you look at a typical customer’s data management infrastructure there are multiple silos of storage and capability that are involved in managing the data. With astronomical data growth rates, IT budgets that are no longer keeping pace, and critical skills shortages the current approach to infrastructure is no longer sustainable given the high acquisition and operations costs. Fortunately, we now have the technology to not only bridge these silos, but eliminate them completely.
At Reduxio, our driving vision is to unify the silos of capability and capacity that exist in a customer’s data management infrastructure, and provide a single converged platform that provides end-to-end management of a customer’s application data, resulting in an infrastructure that is much simpler and cost effective to deploy, own, and manage.
Unifying secondary storage isn’t enough
Solutions that unify the management of secondary data completely ignore the data that is on the primary storage system. The lifecycle of an application’s data must start on primary storage; that’s where an application reads from and writes its data. Excluding primary storage from the data management infrastructure increases operational complexity and costs. In addition, data copies have to be captured from the primary storage and transferred to the secondary storage system using agents or integration with applications, operating systems or the primary storage. At a time when IT organizations are being challenged to deliver ever smaller RPOs and RTOs, this transfer of data puts limits on the minimum RPO and RTO that can be achieved.
Converged primary and secondary data management
Reduxio's converged solution for primary and secondary data management is a single platform that provides a single entry point for the end-to-end management of the lifecycle of application data. The solution provides high performance primary storage, and a set of novel data management capabilities that allows policy based management of data across the primary storage, and external secondary storage attached to the platform. The platform is enabled by Reduxio’s TimeOS storage operating system that provides complete virtualization between the logical view of data that is seen by applications and the physical storage tiers on which the data is kept.
Management of secondary storage tiers is enabled by NoRestore which was delivered as part of the latest release of TimeOS. With TimeOS v3, any block storage system that supports the iSCSI protocol, or objects stores (including Amazon AWS) that support S3 protocol can be used as secondary storage as part of the overall data management infrastructure. Based on simple user policy that can be applied to a volume, NoRestore sends continuous incrementals to the secondary storage which is designated as the NoRestore repository. NoRestore during copy uses NoDup (Reduxio’s dedupe and compression technology) to ensure that only compressed unique data blocks are sent to the repository.
The optimized data movement process allows NoRestore to achieve an RPO as low as 5 minutes with any type of repository. The desired RPO, data retention, and time windows for data transfer are controlled through policy. On the repository, as each new copy is made, NoRestore updates a catalogue which contains the list of all volumes that are stored in the repository and the available recovery points. Any Reduxio system (unconnected with the source system) can attach to the repository, browse the catalog, and instantaneously instantiate a point.
Data virtualization and instant access
Instantaneous access to volumes points stored on the repository is enabled by Reduxio’s virtualization technology. To instantiate a volume a small amount of metadata that represents the point that needs to be accessed is read from the repository. This typically requires less than a minute of time under the worst supported network conditions. Once the metadata is read, the volume is created and can be used by applications before any data is read.
After a volume is instantiated a background process is started to read data for the volume from the repository. Writes from applications are completed immediately with no interaction with the repository. For a read operation for which the data is not available locally, data is read on demand from the repository. NoRestore uses NoDup to minimize the amount of data that is read from the repository - both application initiated and background reads, only unique data blocks that are not available locally are read from the repository.
Redefining workflows with NoRestore
NoRestore (combined with BackDating) allows IT owners to eliminate traditional data protection solutions and architectures reducing infrastructure and operational costs. BackDating provides instantaneous recovery from application and user errors, logical corruption, and malware at the granularity of 1 second. NoRestore provides disaster recovery and increases flexibility during recovery since the shared repository allows for recovery scenarios not available with traditional solutions.
NoRestore allows data to be distributed instantly between remote sites to support a wide variety of use cases including workflows for DevOps and Test/Dev. In the case of sharing production data sets to distributed development teams, once the product data set is captured in a NoRestore repository, it can be instantly made accessible to multiple development sites. Since volume access does not require that any data is available on the target system, the amount of data transferred can be limited to just what is required. With the ability to natively connect to wide range of cloud providers using the S3 object storage protocol, NoRestore can also be used to retire legacy applications to the cloud. An application can be instantly “recalled” and restarted taking advantage of instant restored capability.
Migrating data between systems
NoMigrate allows volumes to be instantly migrated from any storage system that supports the iSCSI protocol to a Reduxio system. NoMigrate uses Reduxio’s unique virtualization architecture and a process that is very similar to the restore process in NoRestore to provide instant access to data after a migration is started. When a volume is migrated using NoRestore, a new volume is created on the Reduxio system that is the target of the migration operation and all the volumes data is accessible and applications can start working immediately. A write operation is always completed locally on the target machine. A background process is started to read data from the source and any reads for data that is not already present on the target machine is forwarded to the source. NoMigrate eliminates the need for 3rd party migration tools or creating copies on secondary storage system to migrate data and applications. In addition application downtime is minimized compared to traditional methods like replication.
Data management starts with Primary
To be able to converge primary and secondary storage management effectively into a single platform requires that the platform provides high performance primary storage. This allows the platform to avoid the use of agents and integrations to capture and transfer data. In addition, there is no resource intensive translation of data formats between different storage platforms. A Reduxio storage system uses a combination of global dedupe and compression, and rapid relocation to deliver the most cost effective primary storage for the most demanding enterprise applications.
Reduxio can provide an enormous boost to application performance while keeping the cost of infrastructure low. This includes a broad range of performance intensive applications including databases, and server virtualization and desktop virtualization. Reduxio systems use a combination of global dedupe and compression (NoDup), and rapid data relocation (Tier-X) to achieve affordable performance. When data enters the system, dedupe and compression and applied inline in real-time and only optimized data is placed in memory cache. Since the data optimization happens before the volume context for a given block of data is assigned in the system, the optimization is global across all data stored in the system, across all volumes and the historical data that is retained to support BackDating. This provides enormous performance gain while reducing the cost of storage. In addition, the unique data relocation engine is media aware and automatically optimizes data movement for the target media type. This technology allows Reduxio to take advantage of new media types as they become available. As an example, you can imagine a system where the first tier of storage is 3D XPoint, the second tier is TLC or QLC 3D NAND, and the third tier is a Public Cloud Object Store.
Reduxio’s first product the HX550 flash storage array is the only flash-disk hybrid system in the market that supports real-time inline dedupe and compression. The HX550 is a versatile and cost-effective solution that has been adopted across a broad range of customers from education and local government to large enterprise customers, in a variety of use cases from supporting critical production applications to distributed DevOps environments.
A storage system with multiple applications
Storage admins looking for a cost-effective way to improve performance while protecting their data are increasingly using Reduxio. The capability to unify a second type of storage and integrate data protection into primary storage reduces the number of components IT teams have to manage, collapses infrastructure silos, and makes everyone’s job easier.
Learn more about Reduxio by requesting a short demo.