Data Explosion: Sipping From the Firehose
The amount of new data in enterprises continues to grow exponentially. In 2020, we created, on average, 1.7 MB of data per second for every man, woman and child on planet Earth. Even the nature of the data we collect has changed over the years, from almost exclusively structured data like tables and databases to today’s unstructured data, including images, videos and, most important, documents. Enterprise data is also becoming fuel for new machine learning and artificial intelligence projects, as organizations use tools to mine data to extract knowledge, give context and provide actionable insights.
How Long Is Forever?
As the amount of data enterprises collect grows, so too does the number of backups that are retained for various purposes. Daily, weekly, monthly and annual backups are often needed to support regulatory audits, financial audits and legal discovery, to name just a few requirements.
Enterprises are increasingly relying on a distributed workforce—whether working from home or across the globe—and as the amount of data grows, users are also working more hours outside of normal business hours, putting more pressure on shrinking backup windows.
The Closing Backup Window
One way to manage shrinking backup windows is to rely entirely on low-cost primary disk storage as the backup medium. This approach is fast, but it becomes way too expensive given the amount of terabytes of disk required for long-term retention of weekly, monthly and annual backups.
The other common data center backup storage approach is inline deduplication appliances. However, they have more challenges than primary disk. First, they are slower for both backup and restore because of the time it takes to perform inline deduplication and rehydrate backups when recovery is required.
The main problem is the lack of scalability with volume. Deduplication appliance expansion is typically disk only, meaning that, although more disk is added, the amount of CPU, memory and network resources in the appliance remains the same. That means the appliance has to deduplicate more data with the same amount of processing power—a physically impossible task. Thus, just adding disk to a deduplication appliance can actually slow down the backup process, shrinking the backup window even further.
This lack of scale-out capability leads to two unwanted outcomes: product obsolescence and the need for forklift upgrades when more processing power is required.
Tiered Backup Storage Meets the Scalability Challenge
Tiered backup storage, with both performance tiers and retention tiers, helps accelerate backup and recovery performance. With a tiered device, when data volume increases, the device can be scaled out, not just up. How does this work? Each tiered system includes not just disk but also CPU, memory and network resources to help keep the backup window at a fixed length even as the data grows. Then, multiple tiered appliances work together to share the backup workloads, so as the data volume grows, each new device automatically connects to the existing network backup topology and takes its share of the backup and deduplication workload. When these factors are considered, it is clear that tiered storage is the answer to the shrinking backup window.
As the industry leader in data center backup storage, only ExaGrid offers tiered backup storage with:
- A disk-cache landing zone that accelerates both backup and restore performance
- A deduplicated long-term repository for cost efficiencies
- A scale-out system architecture that scales as data grows to keep the backup window at a fixed length as data grows
ExaGrid systems never become obsolete, since newer and older devices can work side by side in a single system, each sharing a part of the backup and deduplication workloads. ExaGrid also offers enterprises a five-year price protection guarantee that service plans will never increase more than 3% per year, regardless of the age of the system. Say goodbye to the forklift upgrade.
ExaGrid delivers the best of both worlds by offering both the performance and low cost that enterprises demand. For more information, please visit ExaGrid and review the additional articles and resources on this site.