Backup vendors drone on about “data growth” as often as advertisers invoke “challenging times,” but is there a living person who is surprised by data growth? Backup companies use data growth to shift attention away from the real problem -— the amount of work it takes to manage backups. Traditional backup recovery systems and architectures are antiquated, which is why data growth and new workloads create a crushing management challenge. A modern enterprise backup architecture should scale on-demand to meet data growth, new workloads, and new requirements. Today, customers should evaluate backup solutions not by speed, but whether it simplifies their environment.
The real challenge — too much manual labor
In a world where people automate complex tasks with artificial intelligence and machine learning, backup remains excruciatingly manual. Every day, administrators troubleshoot failed operations, set up new backups, and fine-tune their configurations. With agents, servers, catalogs, and deduplicating appliances spread across multiple locations, backup has more points of failure than a Wile E. Coyote scheme. The fragility of the backup infrastructure is most evident when modifying any part of the environment. One customer described two weeks of agony: “A server OS security patch triggered a database upgrade that required a backup agent upgrade, which necessitated a backup software upgrade that forced a full set of backup client upgrades and a backup appliance upgrade that cascaded to an upgrade of the backup appliance used as our air-gapped replication target.”
As if managing a single backup environment isn’t challenging enough, the average customer runs three different backup applications. Organizations usually adopt additional backup products to handle new workloads that their legacy tools do not support (e.g., Microsoft 365 (formerly known as Office 365), endpoints, and cloud applications). Despite the common purpose of backing up data, running each application requires a dedicated team with unique expertise.
Vendors talk about data growth, but the real problem is work growth for the backup teams.
The underlying technical challenge
Backup teams have to work harder because legacy backup products cannot protect their evolving environments.
Traditional backup tools were not architected for the scale of the modern data center. As organizations have moved from application servers to virtual servers to containers, the number of datasets has increased by 100 times or more. Straight-line backup throughput matters less than dynamic, flexible backup scheduling. The consolidation of files onto NAS servers has created dense environments with hundreds of millions of files. Data speed matters less than scalable metadata management. Since the backup products were not built to address modern requirements, the backup team has to fill in the gaps with manual labor.
Unfortunately, data infrastructure will continue to scale in ways that stress traditional data protection. Kubernetes and containers generate more datasets than virtual servers. Cloud applications like Microsoft 365, Google Drive, and Salesforce.com create more metadata intensive backups than NAS servers. Cloud-native applications use a greater variety of data repositories (e.g. NoSQL databases, object storage) than on-premises applications. Environments are evolving faster than ever, and legacy backup cannot keep up.
The right technical backup architecture
Cloud is not only the future of applications, it’s the future of backup. Cloud is the only place where you can pay for only what you use, near-instantly scale resources, and build applications that scale on-demand. A cloud-backup architecture needs to re-invent deduplicated storage, integrate end-to-end security, and scale on-demand.
A cloud backup architecture begins with revolutionizing deduplicated storage. Systems use metadata, information about the backup, to deduplicate, find files, and migrate cold data, so they need fast access to metadata (but not the data). Backup appliances store metadata and data on one cluster, and they scale by adding boxes. In the cloud, however, there is a significant cost/performance difference between database and object storage and both can be scaled independently. By creating a central meta-database and storing the data in objects, one “system” can manage hundreds of PBs of deduplicated backups on the lowest cost storage available, while still meeting performance service levels.
A secure cloud backup platform protects against internal and external threats. The platform should meet the most stringent security certifications because it dedicates a team to handle patches, pen testing, threat identification, and more. The platform should ensure that only customers can access their backup data because the customer, not the platform, owns the encryption keys. Finally, by being remote from the on-premises location, the cloud backup platform should automatically protect against cyber-threats like ransomware.
The cloud backup can reduce the cost and complexity of data protection with auto-scaling. On-premises, customers must over-provision backup infrastructure, so they can protect the data during a limited backup window and not have to constantly buy and deploy new hardware. The rest of the day, the equipment sits nearly idle. Even if a customer needs to run an urgent restore, most of the backup resources cannot be applied. Capacity planning, no matter how much effort they expend, cannot predict their future. In the cloud, however, a microservices architecture can auto-scale so that you get all the resources you need for a backup or restore when you need them. And, of course, it will auto-scale those resources down when you don’t.
The right business solution — SaaS backup
The cloud backup architecture eliminates most backup challenges. By backing up directly to the cloud, there are fewer components to manage — no backup servers or appliances. With auto-scaling, there is no need to tune schedules, manage appliance capacity, or balance server load. With built-in security, there is no need to retrofit air-gapped backup replication. The right technical architecture solves the real business problem — manual effort to make backup work.
Still, while the right technical architecture is cloud-centric, there is little value in building it yourself. If the goal is to reduce the effort you spend on backup, the right answer is to hand off the work to an expert. Just as customers have adopted SaaS for CRM, email, IT Ops, and HR, they are also moving to SaaS backup.
As you select a SaaS backup vendor what should you look for?
Proven reliability: While most vendors tout the stability of the underlying object storage, it does not mean that their file system is reliable. Find a vendor that can explain their reliability and has a track record to prove it.
Proven security: Backups contain all of your data, so select a vendor that meets certifications including (but not limited to): SOC 2 Type II, HIPAA, and FIPS 140-2. These certifications show that the vendor takes security seriously.
Proven scalability: Every microservices architecture boasts an infinite theoretical scale. Find a vendor that can actually demonstrate that scalability — number of backups, size of backups, number of files, etc.
Proven extensibility: SaaS or not, you do not want to buy more than one backup application. Therefore, find a vendor that has shown it can rapidly and seamlessly add new backup workloads.
With these four factors, you will achieve the business goal — a simple data protection solution, so you can focus on other challenges in your environment.
Data infrastructure today would be unrecognizable twenty years ago, and the biggest difference is not the volume of data. There are more data sources, more metadata, and more locations for data. Even as the infrastructure it protects evolves, most backup solutions still reflect the architecture of the past. Therefore, while organizations automate every other part of their data infrastructure, backup requires even more manual effort to fill in the architectural gaps.
Fortunately, you can leapfrog into the future with SaaS data protection. With a backup architecture designed for the cloud — separating data and metadata, securing the infrastructure and data, and scaling up and down on-demand -— data backup can become the most scalable part of your environment. By adopting a proven SaaS solution, you can shift from backup being your most human-intensive operation to the least. Don’t be distracted by speeds and feeds. Focus on solving the real goal — data backup should just automatically work.