After our Data Protection Launch this summer, somebody asked: “How can you claim to have an open architecture when you still sell hardware appliances?”
Customers fear vendor lock-in. With increasingly rapid technology transitions and “too big to fail” technology companies imploding faster than ever before, nobody wants to compromise their flexibility. That is why, as the world’s largest data protection vendor, we obsess about disrupting ourselves. Wise customers are equally vigilant. Avoiding vendor lock-in is a key value to EMC and our customers.
But it is important to focus on the most critical lock-in points. Backup software, not hardware, is the primary culprit of data protection lock-in. Furthermore, when looking at hardware lock-in, our protection storage is the most open in the industry. Finally, you need an architecture to address all layers of lock-in. That’s why we have the Protection Storage Architecture.
Backup Software – Lock Me Up, Lock Me In
What are the main causes of data protection lock-in?
- Proprietary data formats – Traditional backup applications store data in a closed, proprietary format. You cannot access your data without the backup application that created it. As the old industry joke goes, “You buy my backup app, I own your data.”
- Employee expertise – Backup applications are complex. They become even more tangled because backup teams write custom scripts to bolt on support for protection workflows and configurations that the backup application does not natively support. Unfortunately, what initially seems like job security quickly becomes a hindrance to career advancement.
- Closed Ecosystem – To get reasonable performance, the data source (hypervisor, application, primary storage) demands that you to buy all solution components from them. While there are “alternative” protection methods, they make them so slow as to be unusable. This is most prevalent in NAS environments with closed, proprietary replication software.
There is truth to the old saying, “Backup software is sticky.” Even as backup appliances have displaced the entire tape industry, backup software market share has scarcely shifted in the last decade. Meanwhile, the rise of proprietary replication software introduces an even deeper customer lock-in.
In other words, when it comes to lock-in, it’s all about the software.
Protection Storage – What Is Open?
When it comes to lock-in, people often confuse packaging with architecture. They ask, “Isn’t software-defined storage more open than appliances?” The primary value of any storage solution is in its software. With an appliance, you’re buying embedded software like the Data Domain File System. With “software-only” solutions, you’re manually connecting file system and/or object management software running on extra servers to the storage hardware. Regardless of the model, you’re buying into somebody’s storage software and storage hardware.
As with backup software, most storage lock-in comes from the software. Therefore, the first-order decision is to select the most flexible storage software. In other words, don’t let the packaging distract you from the true discussion of what is open.
What constitutes “open” when it comes to storage software?
Perhaps the best way to answer the question is to think of the worst storage vendor experience you could encounter:
- Dedicated, single-purpose storage – If the workflow changes, the storage system cannot be used. Within its workflow, it is managed different than every other storage in the environment.
- Roach Motel – It is virtually impossible to remove the storage. Once data goes in, it is virtually impossible to migrate or remove it.
Now, think about the ideal protection system in this context.
- Multi-use storage – Your protection storage should support all backup applications, application-direct backups (e.g. databases writing backups directly to the protection storage) and archival (compliant and cost-oriented). It supports storing data in native formats (e.g., VMs as VMs) for backup, instant access disaster recovery and more. As the workloads change, your protection storage software should be a stabilizing anchor on your architecture.
- Easy to remove – Since most protection storage systems retain backups for 90 days or fewer, customers should be able to quickly transition between protection storage systems (assuming the storage is not directly tied to the protection software). Meanwhile, for the customer who wants to transition more quickly, they need high performance access to enable migrations to happen quickly and efficiently.
Data Domain is simple, multi-use protection storage software packaged as an appliance. Are the other backup appliances or “software-only” protection storage solutions as open? You can re-use the hardware, but what becomes of your software investment? What becomes of your data? Are you just extending the backup software vendor’s total control over your environment?
Protection Storage Architecture – End-To-End Commitment to Open
To transform the protection industry from its monolithic, proprietary heritage, we introduced the Protection Storage Architecture. The modular, open architecture frees customers from vendor lock-in.
- Protection storage – As discussed above, this is simple, multi-use protection storage that can be easily switched as appropriate. Furthermore, it should also leverage movement techniques (e.g. replication, Data Domain BOOST) to connect on-premise and cloud storage.
- Data source integration – By integrating with the sources, the protection solution should be able to store data in its native format (e.g. VMs as VMs, not tar images). Furthermore, the management should extend to the data source’s native UI (e.g. vSphere, Oracle RMAN), so that the protection expertise is not locked with one group.
- Data management services – Both data and metadata formats should remain open, so that customers can extend their services without being forced to wait for and use ONLY their backup software vendor’s add-on functionality.
With an open, modular architecture, protection can minimize vendor lock-in. They can easily use different protection storage options (on premise or cloud), different management interfaces, and protection software services.
Today customers suffer from backup lock-in caused by closed, monolithic architectures. They are beholden to their backup software vendor, frustrated by the lack of service, but unable to easily move. The answer is the Protection Storage Architecture – a modular architecture that eliminates proprietary data formats, closed ecosystems, and isolated backup teams.
So, how can we claim being committed to an open architecture while we still offer hardware appliances? Because packaging is not architecture.