SWIFTSTACK WHITEPAPER Build the Business Case: Object Storage for Backup, Archiving and Recovery November 2014
The dramatic and relentless growth of unstructured data is causing headaches for IT organizations and storage administrators. Not only are they looking to keep pace with the need to store more data than ever without busting their budgets but they are also facing a variety of operational, legal, regulatory and security pressures to retain, organize, manage and produce that data on a moment s notice. Data storage requirements are growing more than 50% annually, according to many research estimates, with unstructured data taking up the lion s share of that growth. This is making it harder than ever to back up, archive and recover data coming from multiple storage targets, physical locations, data formats, workloads, storage topologies and compute architectures. While there are numerous ways of dealing with runaway storage growth for backup, archiving and recovery workloads, many organizations are constrained by such factors as cost, manpower, management complexity and performance. Among the challenges: Tape remains attractive for its low capital expenditure impact, but its scalability and performance limitations typically render it ineffective in most enterprise scenarios. Hard disk-based arrays offer capacity, scalability and performance. But they can become very expensive even in scale-out scenarios, and often require over-provisioning of storage to support multiple copies of the same data. Solid-state drives are gaining attention and some support in the enterprise for their performance. But all-flash arrays can fall short on meeting enterprise capacity requirements and are relatively expensive on a cost-per-storage-unit basis. Considering Object Storage for Backup, Archiving and Recovery For storage administrators looking for cost-efficient, reliable, high-performance and scalable solutions for backup, archiving and recovery, object storage is an excellent option. Object storage offers several attractive capabilities for these workloads, including: Nearly unlimited scalability for both capacity and high performance. Data mobility due to referencing objects by ID, instead of file names. Easier deployment due to tight alignment between the storage architecture and essential enterprise applications. Cost-efficient solutions based on affordable SATA hard drives, rather than on more expensive Fibre Channel or InfiniBand solutions. Object storage is now widely adopted in some of the industry s highest-growth and performance-intensive environments, where backup, archiving and recovery can bring an organization to its knees. Cloud service providers, for instance, have embraced object storage in order to keep pace with exploding demand for more and more storage, and with increasing compliance mandates for secure, highly available and reliably recovered data. Page 2
Object Storage for Backup Backup has historically been an important workload for a simplistic reason: It is insurance against unplanned service outages, such as mechanical failure of spinning disks. Of course, most IT professionals will tell you that viewing data backup solely as an insurance policy in the unlikely case of a disk crash is shortsighted at best, and dangerous at worst. IT organizations now know that backup has taken on increased significance for numerous reasons, starting with but not limited to the astronomic growth rates of data especially unstructured data. Certainly, object storage s impressive capacity scalability makes it a great backup medium, but there are other reasons why object storage for backup makes sense. These include: Interfacing with all popular backup software solutions. Fortunately, organizations that want to take advantage of object storage s capacity and management flexibility can feel confident in knowing that object storage solutions are fully compatible with most, if not all, leading backup software without need for software or hardware tuning. Maintaining reasonable and achievable backup windows. With so much data to back up, and much less flexibility in taking down production systems for extensive periods to do that backup, IT organizations need tight backup windows. Object storage is a good fit here, especially when solutions are designed with a multi-tier architecture that allows performance to scale in the proxy tier in concert with the need for capacity scalability. This results in shorter backup windows and lower impact on production workloads. Keeping backup costs under control as data storage increases. One of the biggest benefits of object storage is the ability to eliminate costly and inefficient copies of files, as is typified by traditional hard disk-based solutions. Object storage also helps to keep down backup infrastructure capital expenditures because it can be deployed across commodity hardware. Support for multi-location data centers, including virtualized storage and cloud computing. More and more of today s IT organizations are deploying workloads across multiple sites to improve resiliency and recovery. This means that backup solutions can get unwieldy and inefficient due to the need for replicas across both primary and secondary/recovery sites. Object storage directly addresses challenges associated with multi-location data centers by storing replicas without having to duplicate them over multiple sites. Avoiding overprovisioning of storage to support multiple copies normally associated with other technologies. Hard disk-based backup can require customers to over-buy to meet their backup capacity needs, which also results in expensive and inefficient utilization of additional manpower. Page 2
Object Storage for Archiving and Recovery Of course, as important as backup is for all IT frameworks, it s not enough. Critical data must be archived and recovered in order to support essential business functions across operational, legal, financial and regulatory requirements. Object storage is a cost-efficient, flexible way to support archiving and recovery activities. One of the key requirements for archiving is the ability to automatically move data from primary storage to the most appropriate and cost-efficient medium, depending upon business rules and operational requirements on how often data must be accessed after it has been stored and backed up. Today s object storage solutions support essential storage tiering capabilities in archiving applications, because it combines both cost efficiency and high performance in the same solutions architecture. Object storage clusters, built on commodity storage hardware, are easily scalable to extremely high capacities in alignment with rapid growth of primary storage. Because object storage works with most popular backup software packages, it integrates neatly into archiving frameworks without modification, thus improving management, security and data leakage protection. Finally, object storage is well positioned for disaster recovery and business continuity, especially in solutions based on the software-defined storage model that decouples data from the hardware component of the solution. In scenarios where multiple data centers or recovery sites need to be supported, storage can be accessed from virtual machines quickly and reliably in secondary data centers in the event of unplanned downtime. Considering SwiftStack As object storage continues to gain interest and acceptance by IT professionals and storage administrators, organizations are seeking out and evaluating potential suppliers of object storage to meet the growing capacity, management and economic requirements for backup workloads. One supplier that has committed to object storage as the focus of its solutions portfolio is SwiftStack. It s also important to note that SwiftStack s object storage solutions set is different from that of other competitors in that it is driven by two important principles: a commitment to OpenStack, the increasingly important open source standard, and a focus on software-defined storage, rather than on a hardwarecentric model. SwiftStack s Object Storage Platform combines the OpenStack standard with its own object storage management platform in the company s showcase product, SwiftStack Storage Platform. The platform comprises SwiftStack Controller, SwiftStack Nodes and SwiftStack Filesystem Gateway. This solution is built on the OpenStack Swift Object Storage Engine, a fully open source-based approach, and provides management functionality for both devices and nodes. The SwiftStack Controller supports a browser-based dashboard that enables multiple operators to have role-based access control, while providing single-pane-of-glass management of a distributed storage architecture for backup, archiving and recovery. It enables consistent configuration management across multiple nodes, while supporting reports for capacity planning and storage utilization for chargebacks. Page 4
Another key backup feature of SwiftStack version 2.0 is its support for global data distribution, making it very attractive for multi-site backup targets within a single object storage cluster. This improves performance, which in turns allows for tighter backup windows and improved backup management. Released in the first half of 2014, SwiftStack 2.0 received praise from industry analysts such as 451 Research, which said, The latest software update adds further enterprise-grade capabilities. The research firm added: Interest in both OpenStack and object storage is building; by hitching its wagon to both trains, SwiftStack can take a lead role in the next-generation storage discussion. 1 Conclusion As backup, archiving and recovery workloads become larger, more complex and more important, IT organizations have become severely challenged. They must keep up with fast-mounting data backup capacity requirements, as well as the need for intelligent and highly automated archiving solutions and reliable and instantaneous recovery. Of course, all this must be accomplished against the reality of tighter storage infrastructure budgets. Object storage has emerged as a viable option for backup, archiving and recovery because of its infinite capacity scalability, cost efficiency, reliable performance at hyper-scale levels and management simplicity. Specifically, SwiftStack s object storage solutions, based on the key pillars of standards-based open systems and software-defined storage, are becoming important and attractive elements of IT organizations storage architectures. 1 SwiftStack unveils version 2.0 of its object storage controller, 451 Research, July 2014 Page 5