Friday, October 16, 2009

Oracle Closed World - now closed

It's Friday morning and I'm on my way away from San Francisco after a splendid week of OOW, good guys, a few beers, and a lot of tech talk.

We ran OCW four times from Monday to Thursday, and it was really good presenters we had talked into showing up:

Monday: Jeff Needham on processors and how Oracle runs on them. Opteron good. Nehalem good. A reporter named Kate was present in order to write about OCW. Code: 41.

Tuesday: Jonathan Lewis showing why the crowd were not experts. Ouch. Code 43.

Wednesday: Jeremiah Wilton about the Cloud, and especially the Amazon Cloud. He seems to know a good deal about Amazon. Code 24.

Thursday: Uri Shaft on counting eg. NDV in the optimizer, and some compression theory - and then Dan Norris & Greg Rahn about the Database Machine. Code 42.

And Kate's funny article about OCW appeared in the daily conference newspaper on Thursday. She got all the technical and non-technical stuff right - very impressive! She also gave away the secret location (Thirsty Bear on 661 Howard, upstairs), but thankfully only on the very last day of OCW :).

I truly enjoyed it, and so did several others, so we'll probably do it again next year.

Apart from that, it also appears that the guys from Miracle who were here with me (Morten Tangaa, Jesper Haure, Kaj Christensen, Claus Sørensen) got good things out of the conference.

While I remember it: Thank you to Victoria Lira, Lillian Buziak, and Justin Kestelyn for allocating a reporter for OCW, for managing the whole ACE Director thing, and many other favors that make the conference work.

Wednesday, October 14, 2009

Extra! Extra! Oracle Closed World today.... on Cloud

We had planned not to have any OCW presentations today in order not to steal Larry's audience from his planned keynote, but we're doing it anyway.

It's at 1200 hours, NOT 1300 hours as usual.

More details via text messages later, including todays codeword. If you want text messages from me for the OCW sessions, send me a text/SMS on +45 25277100.

Cloud computing is 'hot'. So is Larry when he talks about it on YouTube. Funny as Hell, actually.

There are at least these two videos. They are partly overlapping, but that doesn't matter- you'll want to see him do this standup routine a couple of times, trust me:

Which is why today, at the secret location, Oracle Closed World will present a couple of guys that know everything about 'the cloud'.


Sunday, October 11, 2009

Oracle Closed World - an underground conference...

I'm here in San Francisco for the Oracle Open World conference along with four other guys from Miracle, the two crazy Miracle Finland guys and some other crazy people - we've rented a couple of big apartments as usual, and are doing work, beer and other essential stuff together.

Last year at Oracle Open World (OOW) my friend Iggy Fernandez, who edits the NOCOUG (Northern California Oracle User Group) magazine/journal, suggested an Oracle Closed World conference, where REAL, TECHNICAL presentations would take place underground in secret locations, using secret passwords, and what have you.

Well, it's here. Monday, Tuesday and Thursdag at a secret location we'll do deep and very technical presentations about various topics. The secret location (which is indeed underground) has the capability to serve beer, by the way.

Let me know if you're interested in hearing more about OCW - email me on or text me on +45 2527 7100.


Things you never wanted to know about SAN's...

Here's some information you will try to forget after reading. It explains why SAN's always cause trouble, why "a firmware upgrade" is really a complete change of an OS and therefor really dangerous (and impossible to plan or test for) and more.

From now on, think of the firmware in a SAN as a whole OS, just bigger. Scary, right?

My question to this very smart guy I know was this:

"Could you repeat what OS'es are used in what SAN's for me? And how many code lines the ExaData is using?"


oh god...that is a huge question... First, Exadata software is small (less than about 100MB of bits)...but that is an unfair comparison to the glut of stuff in a full-featured array... Exadata has Linux underneath it, but then we execute about 98% User, 2% Kernel so really, the only thing we get from Linux is scheduling and I/O... Exadata is small because it doesn't do any of the fat stuff arrays like Clariion do (e.g., snapshots, remote mirror, etc).

Netapp's is called OnTap and it is a heavily developed BSD (Net/1 to be exact). It is huge and full featured as you can tell by how many add on packages it support, but just in protocol provisioning it is huge. Consider the fact that it can support front-end FC yet the LUNS are actually files in the WAFL filesystem! Wild.

Clariion OS is called FLARE and it sits on top of a full Windows distro (XP). EMC NAS (celerra) is called DART which is written from scratch.

HP StorageWorks Clustered Gateway is Linux +hundreds of thousands of very specialized PolyServe code.

EMC DMX OS is called Enginuity...DMX cpus are Power and I have no idea what the origin of this OS is. If I were a betting man I'd bet that it is scratch like DART.

IBM DS83XX is full blown AIX plus more (this is the old SHARK array) in fact, it is a cluster of AIX boxes in there...

I don't know what HP calls the stuff that runs inside EVA ...

as you can is very confusing.