Scaling Oracle in a hyper-growth start up??

Replies to This Discussion

Permalink Reply by Dominic Delmolino on July 28, 2008 at 6:36am

Lem,

These are good questions, and I'm pleased that as a recruiter you're trying to take advantage of this kind of forum in order to become more educated.

I've spent a lot of time lately working with our own recruiters and customers to educate them on database organizations and what skills, roles, tasks and responsibilities are needed to be successful.

Your questions are a good start, but you'll probably need to provide more detail if you'd like to get some concrete feedback. Since you've told us that you're not very technical, it's possible that you don't have more details or aren't even sure what questions to ask. That's ok. Let me see if this helps you.

1700 networks per day doesn't mean much in a technical context -- a database system that adds 1700 records to a table every day isn't something that scares anyone from a growth perspective. I'm being extreme here, but it helps to understand that folks need to know what kinds of data and activity volumes that translates to.

For example, you might add 100GB of data per data and are seeing that grow rapidly (e.g.; it might mean that every day you add even more data and expect to be adding over 1-2TB per day in the near future). Or, you might have 20,000 logins per day and every day that number gets bigger and bigger -- logins translates to activity on the platform. This kind of information about data volume growth and activity growth help to understand what's happening on the platform.

The skill sets that deal with data volume growth and activity growth are slightly different from each other. Data volume growth calls from someone who is good at storage management, storage capacity planning, as well as backup / recovery techniques. Activity growth calls for someone who can do performance management, understands application demands via SQL, and understands how to measure and plan for scalabilty.

It also helps to know what kinds of problems or concerns you have about the platform -- if reliability is a big concern, then you need someone who's good at high availablity and disaster recovery solutions. If performance stability is a big concern, then you need someone who has some architecture and coding background that will help make sure the applications don't abuse the database infrastructure.

Also need to know if you're looking for a single person to "do it all" or are willing to start assembling a team with various roles and responsibilities. Do you need someone to work closely with the developers to make sure the make efficient use of the database (recommended)? I've usually defined 5 roles that describe some of the common database functions:

Database Operator
Database Administrator
Database Developer
Database Engineer
Database Architect

Operators / Administrators typically work in an operations context and focus on reliability. Developers / Engineers typically work in a development context and focus on scalability.
Architects have experience in both domains and can help guide choices.

These roles aren't "official", but I've found them useful when describing database organizations.

I'm sure this post will spark some debate here -- and that's a good thing. Expanding this conversation will help you get some broad perspective.

Good luck and keep asking questions!

Dominic Delmolino
dominic.delmolino@agilex.com

▶ Reply to This

Permalink Reply by Lem on July 31, 2008 at 1:57pm

Dominic,

Thank you for your detailed response to my inquiry. Allow me to provide some more information.

By default, social networks on Ning come with a quota of 10GB of storage and 100GB of bandwidth. This is a network with approximately 5,000 photos or 500 videos. Of course not all networks reach that capacity. Also, on the flip side of that coin, we have a few networks that greatly exceed those quotas. I would imagine that is where one of the interesting database architecture challenges arises and suggests the logical question of "What are the challenges of structuring a database infrastructure that is scalable along a spectrum that included both ends of that capacity spectrum?"

To touch on some of the things you point out...

"It also helps to know what kinds of problems or concerns you have about the platform -- if reliability is a big concern, then you need someone who's good at high availablity and disaster recovery solutions. If performance stability is a big concern, then you need someone who has some architecture and coding background that will help make sure the applications don't abuse the database infrastructure."

All of the above. When considering our network growth rates it's imperitive that we address availability along with scalability. What I've found so far is that there are a bunch of DBA's out there...but very few that have this kind of experience in an OLTP High Aavailability production environment for an internet company. I wouldn't say I'm pulling my hair out just yet, but gettin' there. :-)

I certainly do this topic sparks some interest. There are some really exciting challenges that DBA's face in an environment like we have an Ning, and learning more about them helps me be a more educated person when discussing these things with DBA's like you all. Im sure it's super frustrating for all of you DBA's when you're talking to a Recruiter that has no idea what's really going on in their environment.

Cheers,

Lem Diaz
Operations at Ning

▶ Reply to This

Permalink Reply by Lem on August 21, 2008 at 11:12am

Anyone have any further opinions on this topic? Ning is exploding from a growth perspective and we're currently looking for a Top-notch Sr. Oracle DBA to keep us ahead of that growth curve. Any ideas?

Thanks!

Lem

▶ Reply to This