Solving the Backup Conundrum
Backup can be one of the most frustrating responsibilities of any IT organisation. It can be complicated. It can be time and resource consuming. Sometimes, it can even cause problems. Seldom does it have any immediate measurable value — that is, until there’s a system failure, data loss, a data centre firm closure, a ransomware lockdown, or some other catastrophe. Then, good backups are imperative and critical. Ninety-four percent of companies suffering from a catastrophic data loss do not survive, and without backups, data loss is inevitable.
At one point in the past, backups were relatively easy. Client devices were in a building, on a local network. Servers and storage were in a closet, or a data centre. An organisation had gigabytes (GB), or maybe even a few terabytes (TB) of data. Data theft or ransomware lockdowns were rare. Tape was a reasonable technology for almost every organisation. IT had the access, the skills, and the tools (often a single tool) they needed to keep data stored, secured and safe.
Today, the situation is different for multiple reasons:
- We’re dealing with exponential data growth. From 2010 to 2020, the amount of data in the world increased from 1.2 trillion GB (1,200 exabytes) to 59 trillion GB (5,900 exabytes). Many organisations need to back up petabytes of data.
- Organisations spread their data across the world. For example, 43% of organisations are routinely moving an average of 126 TB from remote offices or the edge to internally managed data centres.
- Almost every organisation of any size uses a mix of hybrid, multicloud, and multi-site on-premises infrastructures. Eighty-five percent of organisations are cloud-first. One out of three organisations are multicloud.
- Backups are broken. Fourteen percent of all data is never backed up and 58% of recoveries fail — leaving business data unprotected and irretrievable in the event of an outage or attack.
- Threats to data integrity and data access are a normal part of doing business. Sixty-four percent of companies around the globe have experienced at least one form of a cyberattack.
- Cost overruns are normal. “In 2020…as cloud usage went up, 42.6% of organisations ended the year overrunning their cloud budgets,” according to a report from Pepperdata. “One-third say they blew past their cloud budget estimates by between 20–40%.”
In other words, we live in a world of ever-changing complexity, and missteps are becoming common. It’s not unusual, for example, for an organisation to choose a public cloud, migrate data and workloads to it, and then start discovering that initial performance expectations and cost predictions don’t align with reality.
As a result, a discipline called FinOps aims to ensure that new technologies — especially cloud — deliver the right capabilities, and the right outcomes, for the right price. Today’s FinOps leaders are scrutinizing every aspect of the IT landscape, and they’re shining a light on backups in our incredibly complex, multicloud world. To experts, that’s not surprising. Organisations know they need better backup solutions — without compromising the need to sidestep cost overruns and unacceptable complexity.
Choosing Cloud for Backup
So, we’ve established that data sets are growing exponentially; complex, costly, hybrid, multicloud infrastructures are becoming the status quo; security is an issue; cost overruns are common; and backups are a problem.
The reality is that the complexity of hybrid, multicloud infrastructures means that data is distributed across many platforms and many locations. A mid-sized organisation might have data over several hundred devices scattered across a few continents. Backing that data up to a single on-premises tape library, or even an on-premises disk storage device, just won’t work.
Traditionally, organisations have tried to solve that problem by adding additional backup platforms in their data centres. And while that approach does give organisations more versatility for performing backups, unacceptable inefficiencies, complexities, and costs arise over time as demands for more storage capacity, longer backup windows, and added features cause organisations to invest more and more money in what ultimately becomes legacy architecture.
Organisations need to sidestep these problems. Becoming cloud-first is one way forward, and that’s already happening as they develop new applications and services that will never live in an enterprise data centre, but will instead reside on compute instances and storage volumes provided by a hyperscale public cloud provider. FinOps teams know that cloud is here to stay and that it offers best-in-class advantages for many requirements.
One of these requirements is storage, and cloud storage is especially suited for backup. More and more organisations are deciding to decommission their on-premises tape-based backup in favour of cloud. Cloud offers a collection of compelling features that make it well-suited for backup. These include:
- Essentially instant, always available expansion
- Ongoing innovation and technology refresh
- High levels of reliability, sometimes resulting in only seconds of downtime per year
- Rich feature sets without the added licensing costs of some on-premises storage platforms
- Reasonable performance for a wide variety of requirements
- Support for storage types ideally suited for backup
- APIs that allow easy access by third parties, including backup software vendors
- Costs that are based on utilisation rather than the capital expenditure of expensive upfront hardware.
The next emerging step is for organisations to accelerate adoption of cloud for backup — and that’s already happening, with many keeping their production workloads on high-performance cloud block storage like Amazon’s EBS and using Amazon’s S3 for backup. This works well at first. However, as organisations grow their backup data sets, need to restore more often, and realise different recovery point objectives (RPOs) for different workloads, some are discovering that backing up to cloud isn’t the panacea they had hoped it would be.
Why? Because some clouds come with complicated and confusing cost models that undermine the value of cloud.
If you look at the pricing page for a leading cloud provider, here’s what you’ll see for one data centre region:
- 14 different storage tier price points, depending on data set size or number of objects
- A very wide range of RPOs, from milliseconds to half a day
- A pricing chart with 85 cells for request and data retrieval charges
- 59 different data transfer charges
- Dozens of charges for management and analytics
In other words, it’s a pricing model that’s almost impossible to understand and incredibly difficult to manage. Imagine trying to manage 40 or 50 backup jobs with different RPOs, the need to migrate data from one tier or storage region, different charges for GET or PUT requests, and added charges for management or analytics. As backup data sets grow and needs change, costs become more and more unpredictable, spiral out of control, and result in budget overruns and executive frustration.
How are FinOps teams trying to address these problems? Often, they’re using a counter-intuitive approach. They’re recognising the true value of multicloud infrastructures and choosing storage-optimised clouds for backup.
Granted, many organisations’ IT departments have more complexity than they know how to handle. But just as organisations know that one server size doesn’t fit all workloads, FinOps experts are learning that there’s wisdom in choosing a best-in-class cloud service for different operational requirements and, particularly where backup is concerned, that some clouds are a better fit than others.
Choosing a cloud that’s optimised for backup may result in performance, functionality and cost advantages. A backup-optimised cloud gives FinOps teams a new tool for reducing costs, reducing cost unpredictability, and reducing cost overruns.
So, what would this look like? An organisation can have multiple clouds, with compute and storage for production workloads. Backup software running in those clouds can push backup data to a backup-optimised cloud, which offers the right mix of features and functionality for the right price. In other words, these organisations can have cloud-to-cloud backup.
What are the ideal characteristics of a backup cloud?
- Open access or support for a widely utilised API for universal access from a wide range of backup software, other clouds, and on-premises infrastructure: For example, S3 is a standard interface and communication protocol — widely used and widely accessible. An S3-compatible data storage repository could be independent of a production cloud.
- The features needed for backup: Massive, on-demand scaling, immutable object lock, reasonable recovery performance, and automated cross-country replication.
- Robust security measures: Anti-ransomware, encryption, powerful access controls.
- No-compromise resiliency: 99.99% uptime or better.
- A streamlined, simplified cost structure: No API or egress charges.
- Limitless scalability: Extend your long-term backups and archives to petabyte scale with a pay-as-you-grow model that sidesteps the expenses of on-premises backup expansion.
- Predictable performance: No “up to twelve hours” RPO limitation; define recovery service level agreements as you see fit.
- Streamlined migration: With no egress fees, migrating data in and out is more cost-effective; migration-as-a-service offerings.
- Immediate access: Data trapped across disparate platforms or on physical media isn’t available for analysis. Businesses who can centralise their backups and use that data for operational insight have an advantage.
Not every cloud offers these features. But some providers offer cloud solutions that are tightly aligned to the needs of backup. With those features, FinOps teams would have a new tool; a storage alternative to optimise operational outcomes and costs. Building on an outstanding cloud for backup, organisations could have a better way to improve data governance and compliance without compromising data security and availability. Also, when data is no longer in your primary hyperscale public cloud, and if your backup cloud has full data mobility, it means that your organisation can use your data with any other cloud service. Migrating out of the dependency of a single provider gives organisations added ability to tailor their infrastructure in accordance with their requirements.
Seagate aims to provide FinOps teams with a cloud that’s right for backup. With Seagate® Lyve™ Cloud, your organisation has a new way to make cloud-to-cloud backup better, solving all the performance, cost, and complexity challenges that come with other platforms.
To learn more, visit us at www.seagate.com/services/cloud/storage.
