A common cloud computing use case is moving backup and disaster recovery functions out of traditional data centers and into public cloud services.
In this blog, we’ll explore how Network-as-a-Service (NaaS) can help enterprises unlock the potential of cloud-based backup and disaster recovery by offering the scalable bandwidth and flexible consumption needed to achieve recovery objectives without breaking the bank.
The need to get real about disaster recovery
Cloud-based infrastructure offers scalable capacity on a consumable basis and easy to access geographical diversity, making it very attractive for backup and disaster recovery. A generation of SaaS-based backup and recovery providers make it easier than ever to access the functionality to take advantage of hybrid and multi-cloud infrastructure. However, it’s important to understand that just taking advantage of cloud infrastructure and software isn’t enough to ensure sound disaster recovery practices.
Gartner analysts have made some pretty devastating observations about the state of IT resilience, including backup and recovery. These comments go far beyond the realm of infrastructure connectivity, but it’s important to note that in their view, many IT organizations engage in what they call “‘disaster recovery theater’ — teams going through the motions in creating plans and play acting with exercises but not actually being prepared for even basic scenarios.” You can read more about this take in the June 2021 research note, IT Resilience — 7 Tips for Improving Reliability, Tolerability and Disaster Recovery by Ron Blair, Belinda Wilson, Mrudula Bangera, and Josh Chessman.
There are many elements that Gartner and other analysts highlight as significant in ensuring that DR is for real and not just for show, including:
- Getting alignment across the business in regards to the scope of the DR initiative
- Settling on recovery point and recovery time objectives that will meet the business’ needs
- Specific scenario-based recovery plans (as opposed to a generic outage that will never happen)
- Detailed plans that can be executed by a broad team
- Building DR maintenance synchronized with change management so you’re always working with current configurations
- Getting the right infrastructural pieces in place, and
- Automating everything
Aside from all of the above, getting serious about DR means facing some uncomfortable realities about the WAN.
Going beyond the obvious with WAN DR challenges
It doesn’t take a rocket scientist to understand that if you don’t have enough bandwidth, you can’t move the bits and bytes to achieve recovery objectives.
Gartner analyst Stanton Cole’s 2021 research note Comparing 3 On-Premises to Cloud Backup Architectures, says it this way: “The WAN is the most critical component in evaluating backing up to cloud services. If the WAN does not have sufficient bandwidth to sustain best-of-breed backup and restore, the overall solution is unlikely to meet requirements.”
So true, but not so simple.
Data movement for backup, restore, and disaster recovery is both cyclical and asymmetric. You have to seed backups with large-scale data movements that require a lot of bandwidth, then you have to maintain a longer cycle of smaller incremental backups. If you ever need to recover, you have a time-critical, large-scale data movement. So yeah, it’s easy to say you need enough bandwidth, but the reality is that enough means different things at different points in this cycle, and the scale of that “enough” has significant economic implications in terms of procuring bandwidth.
The reason for the tension around “enough bandwidth” is because traditional WAN contracts are fixed, long-term affairs. You have to balance the risk of not having enough bandwidth to restore in time against the cost and waste of having a good deal more bandwidth than you may need on average for a period of years.
And the problem gets worse if we go past the primary connection. Long-term contracts mean getting sufficient connectivity to achieve recovery timeframes via truly diverse connectivity is often prohibitively expensive. As a result, many backup network connections rely on relatively low throughput Internet VPN connections. And this same limitation means that conducting realistic exercises is also challenging.
Cloud connectivity adds a further economic wrinkle, because if you’re not using direct connections, you could get bit by significant egress charges when restoring from the cloud.
How NaaS connectivity unlocks cloud DR potential
If WAN bandwidth is the most critical aspect of evaluating backup (and DR) in the cloud, then it’s essential to address cyclical, asymmetric and cloud egress data movement challenges. Fortunately, NaaS is here to meet the moment. With NaaS, you can get bandwidth-on-demand in minutes, on a highly scalable basis. PacketFabric in particular, provides a wealth of consumption options for bandwidth to serve backup and disaster recovery use cases, including:
- Monthly subscriptions for ports and VCs up to 100Gbps
- On-demand Layer 3 connectivity between cloud regions and colocation data centers via Cloud Router, with speeds up to 10Gbps in minutes, and up to 100Gbps connections available on a dedicated link basis.
- Hourly virtual circuits and bursting above monthly subscription VCs, on-demand, for backbone VCs between PacketFabric PoPs
- Usage-based billing virtual circuits
- Aggregate capacity consumption of backbone VCs in 100Gbps increments allowing you to flex VCs up and down at will
If you’re ready to get more familiar with how NaaS can create greater agility for your hybrid and multi-cloud infrastructure, explore our services, request a demo or just register an account and get started today.