Most of you reading this have had your hands in the IT world for a long time: maybe a decade or more.
You've probably noticed that certain events occur along the way that signal a change in approach: it becomes clear that it no longer makes sense doing things the old way, and there's a new (and better) model at hand.
TCP/IP changed the way we did networks: goodbye to SNA, BSC, X.25, even OSI. Shared storage arrays changed how we store data: no more per-server dedicated storage. Cloud is changing the way we think about producing and consuming IT services.
And, of course, VMware has permanently changed the way we thought about servers.
Yes, it might be hard to imagine our IT infrastructure world before those developments, but it wasn't all that long ago.
But there's a pattern that's worth taking note of: once the new technology is in the marketplace and has demonstrated its advantages, the IT world immediately splits into two groups.
The first group recognizes that there's been a fundamental shift, and starts working to incorporate the new thinking in their environments. They know where they want to be down the road.
The second group remains either unaware of (or perhaps opposed to) the new approach, and thus resolutely keeps with well-established practice. And, of course, as an industry we do like to argue: members of the two groups will inevitably debate the perceived merits of their chosen approach.
As a familiar example, consider what happened in the server world over the last decade.
You'll occasionally meet the pioneers who were using early versions of VMware's GSX product way back in 2003 to virtualize modest servers. What was once esoteric and exotic is now common practice.
Such is now the case with many Oracle environments, and the database administrators (DBAs) who look after them.
With the availability of vFabric Data Director 2.0, there will shortly be two classes of Oracle DBAs -- those that recognize the shift and are moving in that direction, and those that aren't.
History does repeat itself.
Take any IT thing with a physical persona; virtualize it; and -- at once -- it becomes much more efficient to produce, manage and consume.
That's obviously true when it comes to server images; it's also true when it comes to networks, user desktops, storage --- and now databases. Oracle databases, in particular.
Simply put, vFabric Data Director virtualizes databases, making them easier to produce, manage and consume. I've written about the product before -- it's one of those cool, potentially-change-the-game-for-IT technologies that doesn't get a huge amount of attention.
The product was promising in its 1.0 form -- but not quite there for the majority of use cases. With the release of 2.0, though, vFabric Data Director now demands wider attention.
How To Explain It?
A good starting point is by drawing an analogy on how vSphere and related tools handle server images, since most readers will be familiar with what it does.
Server images are really just a library of templates. Any one of them is easy to patch or update. They're easy to invoke. Make them highly available with just a click if needed.
Server images are easy to clone. They're also easy to make consumable by others without the server admin having to do too much. They're easy to monitor and manage in the aggregate. And so on.
If you're a long-time server admin, think back a decade on how you managed x86 servers prior to VMware. Now think about what your world looks like today. Big difference, yes?
Now, replace "server images" with "Oracle database images" and you'll quickly understand 90% of what vFabric Data Director now does. Instead of the server admin, it's used by the database admin. In a nutshell, vFabric Data Director makes it vastly simpler for DBAs to manage a fleet of Oracle database images, and offer them up as a virtualized service for others to consume.
It has obvious potential to transform the role of the Oracle DBA from being consumed with rote tasks to exercising their unique skills. And, of course, do a whole bunch of good for the organizations that use a lot of different databases.
The more database instances (production, test, development, decision support, etc.) you have sloshing around in your IT environment, the greater the impact. It's not unusual to meet customers with many thousands of database instances at hand -- in these cases, the potential impact is very great indeed.
Database-As-A-Service?
Yes. Think for a moment -- how many ad-hoc requests for a database come through your shop in a month? What's the workflow and level-of-effort involved from the time the request is made until the database is turned over for use by the requestor? How much time elapses while all of this is being done? Brace yourself: how much unique value exists in performing these routine tasks?
That's the rationale for vFabric Data Director: automate the stuff that doesn't matter so you can focus on the stuff that does.
Thinking at a higher level, more than a few IT organizations are restructuring their operations to deliver services vs. silos. IaaS, PaaS, et. al. It's hard to imagine there isn't a strong need for a standardized way to automatically provision new database instances on behalf of requestors.
They've Made It Easy
There are plenty of online assets explaining what the product does: this presentation, what's new in this release, an easily downloadable version, plenty of documentation, etc. If you've got a bit of time and inclination, it's all there for you to evaluate -- today.
Even if you're not entirely sold yet on the idea of virtualizing your smaller and less-critical Oracle databases, perhaps you owe it to yourself to figure out what's going on -- and form your own opinions.
If nothing else, that's exactly how VMware's first server virtualiztion products got popular way back when ...
Game on!
Disclosure - EMCer here...
vFDD does that and much, much more. You nailed the high level takeaway - but it also abstracts the database access, creates a tenancy model, pools database resources, creates an API model that can be used programmatically, rich replication capabilities, accelerates ingest - oh, and also includes an optimized Postgres distribution for the times you don't want/need full blown Oracle.
Great for any customer - I would highly encourage checking it out.
For EMC customers, on top of that, did I mention that Isilon can help with database ingest from RMAN backups, and vFDD has native embedded DataDomain Boost integration? :-)
Check it out man! http://virtualgeek.typepad.com/virtual_geek/2012/06/vfabric-data-director-making-ingest-and-backup-better.html
Posted by: Chad Sakac | July 16, 2012 at 10:32 AM