More and more folks are first protecting their data (backups, archives, etc.) to disk — for several good reasons, including SLAs for data retrieval/restoration/recovery, deduplication efficiencies, etc. But as definitively as disk is the preferred “protection storage” as a second tier, the third tier is far less defined.
For some, the deterministic factors are BC/DR preparedness or OpEx efficiency, which can point folks to the cloud, especially when the cloud is more than just ‘storage’ but also includes the capability to fail-over. But even then, which cloud? A backup-service (BaaS) or a storage service (extending on-premises backups) or an infrastructureservice (hosting a virtualized backup appliance)? Even when ‘cloud’ is an option, there isn’t a single outcome.
(Disclaimer: ‘cloud’ isn’t technically a medium but a consumption model, where the service provider actually is using disk or tape for the actual storage).
For others, length of time, cost per TB or a compliance mandate point folks to tape. In many cases, those folks are rediscovering modern tape as just that: modern. Cartridge durability and performance ratings prove that today’s solutions are not your daddy’s (or grand-daddy’s) tape solutions of old.
And to be perfectly clear, today’s IT professionals should not be thinking "tape or cloud" any more than they used to think “backups or snapshots” — each have their place as complementary methods.
Here’s a video covering the idea of disk, tape and cloud (not or):
To help ESG clients understand these trends further, we've recently kicked off two research projects:
ESG Data Protection Cloud Strategies (DPCS) – covering which cloud(s) and why, including Backup-as-a-Service (BaaS), Infrastructure-as-a-Service (IaaS), adding cloud-storage to on-premises backup/archival solutions, and backup preferences for Software-as-a-Service (SaaS).
ESG Long-Term Retention Strategies (LTRS) – covering archival trends, regulatory mandates, long-term retention tools (backup vs. archive mechanisms) and tape/cloud repository preferences.
Announcer: The following is an ESG video blog.
Jason: Hi, I'm Jason Buffington. I'm the senior analyst at ESG, covering data protection. One of the more fun debates in data protection storage is disc, dedupe or simple, or tape, or cloud. Let's table the question of whether dedupe should be done in an appliance, physical or virtual, or within the back of software, but just talk about the media. Disc, tape, cloud. Cloud is actually not a medium. It's a delivery mechanism with a different economy model. Cloud storage providers use disc for sure, and a few are also offering nearline solutions, some of which are based on - wait for it - tape.
If we suspend that entire tangent argument on "cloud isn't really a medium type, but a delivery model," then we can go back to the question, which medium types are people using in data backup? Let's look at the data from ESG's data protection modernization trends, where we asked about disc, tape, cloud, and every combination in between.
The most common scenario continues to be what I call the gold standard of disc to disc to tape, D to D to T, where you back up from primary disc, production, to secondary disc for efficient dedupe, fast and granular recovery, etc., and then to tertiary tapes for long-term retention. About one in four organizations do D to D to T today. In second place is a veritable tie, around 16 to 17% each, with a very interesting contrast. Sixteen percent of organizations are using the newer gold standard of disc to disc to cloud, D to D to C. Same rapid granular recovery to disc, and then to cloud for tertiary retention.
In my mind, the tradeoff here is around regulatory compliance, as to how long the tertiary retention is on cloud versus tape, and scale. Because how much data you have to retain for a long period will, or should, affect your cloud versus tape discussion in long-term retention. Maybe that's why the other contender for second place, at 17% of orgs, is disc to tape to tape, meaning that tape is used for on-prem recovery and then some tapes go off-prem for retention, BCR preparedness, etc. And that's kind of funky, if you think about it.
The second most common method of data protection used today is an all-tape solution. Coming in as a third band of choices are the all-disc options. Disc to disc for fast on-prim protection, but no tertiary or off-prem retention - 14% of orgs. Disc to disc across a WAN to presumably a third disc, so it's an all-disc everywhere solution, 11%. And disc straight across a WAN to another disc, like for a centralized backup of branch remote offices back at headquarters, that's 6% of orgs. Disc to tape, old-school. Regular tape backups. Also 6% of orgs. And bringing up the rear, for today at least, disc to cloud, 4%.
I can make a case for almost every one of these solutions, and that's my point. Your recovery goals and your retention requirements and your economic considerations should all come into play here. Me personally? I love cloud backups for remote offices. Though depending on your recovery needs, I could see bringing the data back to headquarters if that's who's going to use the data during a crisis. I cannot say enough goodness about cloud protection for endpoint devices. Just do it. But ensure that you have IT oversight and you're not using a consumer backup as a service. And I think the cloud is very compelling in a BCDR solution where you're not only using cloud-based storage, but also cloud-based compute in order to fail over. Said another way, why BaaS when you can DRaaS?
I'm a fan of cloud-based services and data protection. I'm also a fan of tape, especially if you're in an industry that requires retention or if you're storing data for regulatory or operational purposes for greater than two years. And by the way, any way that you look at it, disc ought to be the primary means of protection recovery before you go to whichever tape or cloud or tertiary disc that you might eventually retain data in, because it is really hard, if not impossible, to reliably meet today's SLAs without a fast on-prim copy of data to recover from.
If we add up the different combinations in the research, the data tells a similar story. Seventy-three percent of organizations use disc as their first tier of recovery, and I would expect that number to gently continue to rise in the future. Sixty-nine percent of organizations use something other than disc in their overall strategy. Forty-nine percent, nearly half, are using tape, where 20% are using cloud. Twenty-three percent of organizations exclusively use tape. Nearly one in four. Four percent exclusively use cloud today. Do I expect the tape number will recede some and the cloud number will rise? Maybe, but not as much as you might think. In the case of tape's decline? Yeah, I expect that some of those that are using tape exclusively will evolve to a disc plus tape model for more efficiency and for tightening SLAs, but it's those same goals.
Efficiency and SLAs are what will likely keep pure straight-to-cloud solutions relatively low. I could do a whole 'nother podcast on D to D to C and the myriad ways to mix those parts between backup and service architectures, hybrid infrastructure architectures, and cloud storage augmenting on-prim backups. But that's for a different day. In the meantime, disc and tape and cloud each have their place around rapid recovery, reliable retention, and data survivability and agility, respectively. So if those are your goals, then arguably, all three media should be part of your strategy for one scenario or another.