Leveraging Machine Learning To Improve Backup Automation Monitoring 

The world of backup and storage management has historically been a manual, reactive endeavor. After all, a functional area that’s focused on backing up data successfully and reporting on those successes involves repetitively looking at prior day, week, or month activities to ensure compliance with disaster recovery and data retention policies.

However, with today’s drive toward automation in a world of data proliferation and resource constraints, this manual, reactive approach simply does not cut it. The sheer volume of data being backed up, coupled with unstructured work-flows spread across servers, geographies, and business units to address a host of needs means the days of responding to failures and looking backwards must come to an end. 

Data protection teams will never be able to hire enough people to ensure that data can be properly protected. 

To get ahead of these challenges Bocada is expanding backup monitoring and reporting automation to include machine learning tools. We’re preparing to give our customers—everyone from backup administrators and systems architects to data auditors and storage teams—the ability to be proactive and anticipate data protection challenges before they happen. Backup teams will be empowered and critical data will be protected. 

Why Machine Learning In Backup Automation

We’ve been in the business of automating backup monitoring and reporting for nearly twenty years. Why add backup failure predictions too? We have two really good reasons. 

Staying In Sync With Customer Needs

As a customer-driven organization, we make a point of listening when our customers ask for features and functionality that would make their work lives easier or simpler. It’s become clear in working with backup and storage teams across industries and the globe that being proactive is the next step on their automation journey. However, between fighting everyday backup failure fires and staying one step ahead of ever-growing backup environments they are barely treading water. They are asking for an oar to paddle forward. We’re building it. 

Staying In Sync With IT Trends

As a leader in the data protection space, we keep our eyes open for technology trends that can benefit backup and storage professionals. The rise of machine learning and predictive analytics is one such opportunity. Encompassing a host of technologies that let systems continuously mine growing historical data, interpret the interrelatedness of that data, and anticipate future events, predictive analytics is tailor-made for a field overflowing in terabytes of backup performance data. 

What Predictive Analytics Delivers In Data Protection

Anyone in our field knows that backups fail for a wide range of reasons, and finding plus fixing the underlying cause of those failures is at the heart of strong data protection practices. What if we could make it easier to not just find but predict those causes with increasing certainty? That’s what bringing machine learning to backup performance is all about. Consider a few potential scenarios:

• Storage capacity forecasting: Hitting backup capacity limits is a common cause for backup failures. So focused on monitoring individual backups and fixing issues, admins often don’t see storage running out until it’s too late. Using predictive analytics, we can predict trends in storage usage relative to capacity, alerting admins of potential storage issues before they result in failures. 
• Backup window failure prevention: Backups must finish within an allotted period of time to be considered successful. What if admins didn’t have to wait until they walked into work the next morning to find out a critical backup didn’t complete within that period? With predictive analytics, systems can assess the amount of data backed up, forecast the incremental amount of time needed, and alert admins if backups will likely fall outside of their success windows. 
• Storage cost forecasting: As organizations migrate their data protection activities to the cloud, storage costs will be a larger part of discussions than storage capacity. How will organizations anticipate their storage cost needs? Predictive analytics can track storage usage trends, forecast future usage, and alert teams when they’re likely to hit desired spend limits.
• Predicting data spikes: How easy is it to identify unusual spikes in backup capacity usage today? With the inputs of historical usage trends, predictive analytics will be able to identify when usage is expected to surge, alerting admins of issues before they even happen.  

This is just a small sampling of issues admins encounter when managing backups and storage. But, it shows how machine learning can save teams hundreds of hours every year by preventing the need for ticketing, troubleshooting, and backup maintenance activities. 

IT Operations Needs Backwards- And Forward-Facing Analytics

Will Bocada continue to give backup and storage administrators an automated, efficient way to monitor and report on historical backup performance? Absolutely! We know that the need to show compliance with government regulations and SLA goals will continue to grow. 

However, we’ll also be giving teams ways to efficiently anticipate issues before they impact workloads, compliance, and SLA targets. Backup environments are simply getting too big and too complex to effectively manage without more automation. 

As we move into the back-half of 2019 and beyond, get ready for releases featuring machine learning and predictive analytics. We’re confident it will be a critical element in helping backup administrators eliminate reactive, manual processes and shift to automated, proactive guardianship of critical data.