Scaling Backup with Cohesity on Cisco UCS

By Nathan Antons • April 23, 2019

Meeting the demands of massive data growth continues to be a challenge facing IT administrators. According to a recent IDC study, the global amount of data will grow from 33 zettabytes in 2018 to 175 zettabytes by 2025. Addressing this over fivefold growth in data will require modern solutions across the data center, particularly when it comes to backup and recovery.

Several key challenges face IT administrators when scaling backup capabilities for their organizations, including:

  • Scaling capacity without disruption– Legacy solutions can require forklift upgrades when they reach their capacity limits. This leaves administrators with a costly choice to disrupt operations with major migrations to larger systems as their data grows, or to overprovision backup with unused capacity before it is needed.
  • New locations for data– Data is now located, and can move between, edge, core, and cloud environments. IT organizations may not have expertise at each edge location to effectively provision and manage local backup appliances. Incorporating cloud into backup solutions can require disparate management systems and bolt-on gateways. Moving duplicate data to the cloud can also introduce unnecessary costs.
  • Degraded Performance– Frequent snapshots can degrade virtual machine performance. Data sets containing a large number of snapshots can slow backups and limit the ability to meet strict SLAs for Recovery Point Objectives (RPOs) and Recovery Time Objectives (RTOs).

Cohesity has pioneered a software defined, scale-out backup solution that modernizes backup and provides key capabilities to enterprise IT organizations to address these challenges. These capabilities help organizations address the very real problems of continued data growth. By partnering with Cisco, we place Cohesity software on world-class server infrastructure. As a result, our shared customers can benefit from modern backup and recovery on their preferred, trusted server platform.

Scaling capacity without disruption

Cohesity software on Cisco UCS offers a modern scale-out architecture for backup and secondary data. With the Cohesity scale-out architecture, you can start with 3 nodes and grow non-disruptively to an unlimited web-scale, while avoiding the challenges of either overprovisioning the capacity of the hardware or growing through forklift upgrades. Maintenance for Cohesity clusters is also non-disruptive, for software updates and the replacement of nodes. Additionally, data is redistributed so that scaling capacity doesn’t cause uneven data across nodes in a cluster. Capacity growth can be managed with granular options through a broad selection of certified servers. Node capacity ranges from 24TB of capacity with the UCS C220 servers, to 420TB in capacity per node with UCS S3260 servers.

Backup at edge, core, and cloud

Cohesity on Cisco UCS will also help you grow and manage backup capacity off-premises, as applications grow beyond core data centers into multiple cloud and edge environments. A frequent challenge of placing appliances at edge locations is to place trained staff at disparate locations. Another challenge is the seamless, efficient movement of data between locations. Cohesity DataPlatform Virtual Edition is an ideal solution for edge and remote office deployments. The backup solution can be deployed quickly as a virtual appliance into existing vSphere and Hyper-V environments, offering the benefits of the core data center solution to edge locations. This solution can be deployed centrally by specialized staff, reducing the maintenance requirements at remote offices. Because a new physical appliance is not required, consistent infrastructure management and spares with platforms such as Cisco UCS can be used across core and edge locations. Cohesity solutions offer global, variable length deduplication across applications within a cluster. This leads to greater efficiencies when sending replicated data that has been deduplicated from edge sites back to core sites, or to clouds. Cohesity DataPlatform features native cloud integration, so seamless replication between core, edge and cloud is offered with this efficient movement of data.

Performance at Scale

As data grows, IT administrators will also want to scale performance with backup activities. A large amount of snapshots can slow backups, which can prevent backups from completing within their window. In legacy storage, snapshot updates form a linked chain, with each link containing the changes from the prior snapshot. Performance overheads increase proportionally to the growth of the chain. Cohesity features SnapTree, which is unique in its ability to support unlimited, frequent snapshots with no performance degradation. By arranging a proprietary tree structure of pointers to snapshot data, Cohesity SnapTree radically reduces performance overhead normally associated with frequent snapshots.

Frequent snapshots can also degrade the performance of virtual machines, and interrupt operations while snapshots occur. Since Cohesity software can maintain a virtually unlimited number of snapshots, administrators are offered the ability to offload frequent backups onto a scalable platform. Cohesity maps each incremental backup to an incremental snapshot, with each snapshot available as a fully hydrated copy. With SnapTree, each incremental backup becomes instantly available as a full image, eliminating the need to perform periodic full backups. We integrate with Cisco HyperFlex snapshots to further reduce impact to operations with deployments of Cisco hyperconverged infrastructure. Cisco HyperFlex snapshot integration offers several benefits including reduced stun times for app-consistent backup, agentless app-aware backup, and improved backup performance.

Cohesity also offers that ability to quickly recover a growing number of virtual machines. Cohesity’s instant mass restore enables IT admins to recover not just files but hundreds of virtual machines instantly, at scale, to any point in time. This powerful capability enables a scalable number of virtual machines to meet strict business SLAs with RPOs in minutes and near-instant RTOs.

Learn More
Modern backup and recovery solutions are necessary to face the challenges presented by massive data growth. Our software defined solution offers unique capabilities to scale non-disruptively, manage backup at core and remote sites, and to maintain performance at scale. We work closely with Cisco to place our software on trusted, world class server platforms for a comprehensive solution.

Learn more about our solutions with Cisco here.

madison