The other issue is not a Linux issue per se. Linux users tend to rely on Open Source software. In general, Open Source software lacks the data protection features that proprietary commercial software has. You see this especially in databases. They're getting better, so much better that companies like MySQL are licensing or selling some of their technology to mainstream companies like Oracle.

Vendor support for Open Source applications is also lacking. Third-party data protection applications usually don't support the LAMP infrastructure. This is particularly true of e-mail and continuous data protection. All CDP vendors support Microsoft Exchange and most support Lotus Domino. That's it. Nothing Open Source.

Major vendors focus on products where they can get support from other software vendors, not a nebulous community of programmers. And this is getting worse as consolidation reduces the number of major data protection software vendors.

The lack of underlying data protection features has a wider-ranging impact on Open Source users. Many applications are built on layers of Open Source software so there's a domino effect.

LWM: What can Linux users do to better protect mission-critical enterprise data?

TP: Sadly, the most important thing to do is stick with Red Hat and SuSE. That's the only way to get support for advanced data protection features. You also have to stick with closed server applications. That means Oracle, Informix, DB2, Lotus Domino, and similar applications. It's lousy but that's life.

Remember that data has value. There's a cost associated with its destruction. It's no different than losing a physical asset, just harder to replace. You can always buy a new server but the same doesn't hold true for your customer data. What you save with a LAMP infrastructure or by using an unsupported distribution may be lost many times over when you lose your data.

The only way around this is to look for data protection appliances or solutions implemented in the storage system itself. For example, rather than use a host-based remote copy, look for a remote copy appliances or a remote copy implemented in the storage system. These can cost more and may be overkill for many applications but they should support even unusual distributions. But that still doesn't buy you application software support.

Other than that, Linux system administrators have to do what everyone else does. I suggest looking to policy-based data protection from the start. Develop policies first and then develop data protection strategies to meet those policies. There's a tendency to do things the other way around. You buy replication software and then look around for ways to use it. Policy and process should always come first.

LWM: What should the Linux community expect from vendors regarding data protection?

TP: More than anything else, Linux users have to push vendors to support the distributions and applications they use. If you reply on LAMP for your infrastructure, then get data protection vendors to support it. The best way is to demand it through the people you buy servers and storage from. If you just spent a boatload of money on IBM blade servers using Linux, IBM has to help insure that you can get advanced data protection features. It's incredible that the CDP product from Tivoli, IBM's management software division, doesn't support Linux and IBM's blade servers do!

Linux users have to push systems vendors to insure support for their environments. That includes Open Source applications.

LWM: What should they expect from the Open Source community?

TP: In a sense the best thing Linux users can do is to help themselves. That means developing industrial-grade Open Source backup, file replication, CDP, and host-based remote copy tools. Not only would that create an alternative to the relatively expensive offerings on the market today, but it might tweak vendors into supporting the LAMP environment better.

Conclusion
There are big changes going on in data protection. Practices such as policy-based data protection, data lifecycle management, and information management are helping system administrators and IT managers make better decisions about the value of their information and what resources should be applied to it. New tools and technology, especially continuous data protection and disk-to-disk backup, are helping to meet the more stringent service levels that come with the 24x7x365 environment. LAMP users are at somewhat of a disadvantage since many of tools don't support their environment fully. Pressure from the Linux community and Open Source initiatives in data protection can help change this. Linux is mission-critical and its data protection tools and practices should reflect that.