Why Kasten Matters to the Data Protection Industry – and to Veeam

This week, Veeam announced its purchase of Kasten. To appreciate the importance of that, you have to look back at how IT data protection cyclically evolves.

The Data Protection (and IT) Evolutionary Cycle

About every ten years, a few elements of IT history typically repeat themselves:

PHASE 1: The production platform of choice for IT fundamentally and architecturally changes. It starts as a niche architecture that meets a nascent need, but as early adopter organizations start to see the possibility, it becomes a mainstream IT platform – and seemingly overnight, everyone is talking about that platform as the inevitable future.

PHASE 2: When the production platform starts to shift or expand, the defacto marketshare leaders in data protection typically try to apply their last-generation approach to backup to the new model. It appears to work on the surface, because they are thought of as industry leaders and they are using language that their core backup buyer is comfortable with like “agents” and “storage snapshots.”

  • Part of why those backup leaders take that approach is because their forward-leaning customers are demanding “any” kind of answer to assure them that what they have been using and intend to use in their present-mode IT is still the right choice. So, those customers want to be assured that their vendor will continue to be relevant and leading-edge.
  • The other part is because those vendors are mainly focused on retaining their current customers and expanding through those organizations, they tend to minimize investments in R&D around what will happen next. This creates an opportunity for startup disruptors.

PHASE 3: One or a few small innovators recognize that the production platform will require a reimagination of how data protection should be done. It is often with the acknowledgement that the production platform lacks the API’s or other “plumbing” to make data protection universally achievable, so the innovators have to do more of the technical heavy-lifting to accomplish the kinds of backups that the mainstream will invariably demand.

Those innovative little backup solutions are subsequently deployed for the new platform as a “complement” to the overall IT backup strategy and the incumbent mainstream backup tool. And this works fine, until the overall IT platform of choice starts to shift to the new platform.

Two Options on What Happens Next

Option 1 – Market Normalization through APIs

In some cases, the next gen production platform vendor recognizes one of the adoption blockers for mainstream users (versus early adopters) is the lack of foundational IT services like being back-up-able. As such, after the first years, the platform creates new methods for backup software vendors to interact with the data. We saw this when Microsoft added VSS nearly ten years after NT was first introduced and again when VMware added Storage APIs for backup in v5. At that point, legacy vendors try to catch up by using the APIs, while the innovator startups try to capitalize on their early-mover advantage to maintain dominance.

Option 2 – The Tail wags the Dog

The legacy IT platform is eventually usurped by the new IT platform. In doing so, the little backup solution for the new platform eventually becomes the organization’s primary backup, since the platform being protected has become the organization’s primary production platform. This can be accelerated if the little/niche backup solution also intentionally expands its use cases from just the new platform to also include protection of the legacy platform(s).

Where do we go from here?

Everything described above is historical from AS400 to NetWare to Windows to Virtualization (from the mid-1980’s to 2020), but there is something different this time. In the past, mainstream IT had one preeminent platform of choice – e.g. when one embraced Windows Server, it came with the assumption that NetWare would be phased out of the datacenter. Similarly, the measure of Virtualization is whether a datacenter is 70%, 80%, or even 90% virtualized (with very few and individually justified physical platforms remaining).

This time around, there is not a single or inevitable “hero” architecture for Hybrid-IT

  • Within the datacenter, most organizations have given up on virtualizing “everything”. Instead, there is a recognition that some workloads ought to be on standalone metal – and for “everything else” (on-premises), virtualize it. Moreover, almost no one believes that the datacenter itself will go away any time soon. Instead, Hybrid-IT is the desired endstate, not just the transitional period between on-prem and the eventual and universal cloud of the future.
  • As for “the cloud,” there isn’t one – there are many and will be many. The data protection ramifications of that are significant, because each kind of cloud architecture requires a different approach to backup and recovery. Just like you should not back up on-prem VMs like you back up on-prem physical servers, you should not (and cannot) back up IaaS VMs the same way as SaaS applications or containers.

This is where Kasten comes in. Kasten recognized that legacy approaches to protecting containers would fall short and so they reimagined what was necessary to back up that nascent platform in its early days – and as containers as a concept and Kubernetes as the predominant approach emerged, Kasten’s vision and strategic significance grew with it – much like Veeam rose while VMware gained prominance over a decade ago.

As an IT vendor, particularly in the data protection space, it is your responsibility to not only continually delight your existing customers, but to innovate (through building or buying) for what customers will want next. Each IT production platform had its early innovators in data protection:

  • In the 90’s, NetWare was best protected by Palindrome and ARCserve (spelled correctly then, later purchased by CA, then spun back off)
  • In the 2000’s, Windows was best protected by Backup Exec and Commvault Galaxy/Simpana
  • In the last decade, VMware is best protected by Veeam

It is worth noting that for each production platform generation, there are early innovators. Some become the dominant backup platforms for their era (listed above) and the rest are bought for their IP or eventually go out of business for not reaching success velocity (due to technology or go-to-market). Looking ahead, for a data protection vendor to be successful in the next era of “the cloud”, they’ll have to address:

  • IaaS backups across all three of the hyperscalers (AWS, Microsoft Azure, and Google)
  • SaaS – with Office365 being by far the most strategic and pervasive
  • And Containers, which is relatively synonymous with Kubernetes at this point

When you modernize production,

you must modernize protection.

Along with that axiom, two lessons that IT history has taught us are:

  1. You cannot (or should not) back up tomorrow’s IT with yesterday’s mechanisms.
  2. The leader in data protection is determined by who has the vision to reimagine backup for what’s next, regardless of how they backed up what came before it.

It’s those last truths that make Veeam’s investments in backup for Office 365 and AWS/Azure over the past few years so important – and why Veeam’s purchase of Kasten is so significant.

This article was originally posted on LinkedIn, with the following context intro:

I’ve been in the #DataProtection industry for 31 years, starting as an IT Pro swapping tapes, later as a channel partner deploying new #backup solutions, but for the most part within backup or replication software vendors that were (fortuitously for me) disrupting their respective markets. Disclosure: I currently work for Veeam.

Solution Diagrams for Kasten (left) and Veeam (right) in 2020
Solution Diagrams for Kasten (left) and Veeam (right) in 2020

Leave a Reply