Posts Tagged ‘Oracle’

Technical details on the Exadata Storage Server

By Paul Vallee September 24th, 2008 at 5:59 pm
Posted in Oracle
Tags:

On Darrin Leboeuf’s advice, I loaded Kevin Closson’s blog, and sure enough, he had something ready to publish.

It must have been KILLING Kevin to keep this a secret. It must be a huge load off to publish this thing.

Anyway, Christo has been assigned to study this in detail and digest it overnight. Expect some good analysis tomorrow.

So, here is Kevin’s post publishing some technical details.

Cheers,
Paul

P.S. I am sure this is the future direction of storage intelligence. The fact that Oracle is setting the bar to only formatting half the disks in order to satisfy the IO saturation of the bandwidth will set a new bar. Do you know how hard we work to convince customers to do this (and fail!?) That, and the connection to BAHD and the problems I laid out in that article are obvious, this approach sets a new bar and addresses all of those issues.

Christo, by the way, is willing to bet this is a full-blown Oracle instance running on each Exabyte Storage Server. Interesting idea. From a manageability point of view, this is mind-boggling but possible.

The Oracle Database Machine, In Partnership with HP.

By Paul Vallee September 24th, 2008 at 5:17 pm
Posted in Oracle
Tags:

Notice to readers:
This is an excerpt of my liveblogging of the Keynote where the Oracle Database Machine was announced.

It is a mix of my comments in real time, and my quotes from things Larry Ellison said that I felt were worthy of mention.

You may be interested in reading more about Oracle’s Exadata platform. I would suggest taking a gander at Oracle’s product page and also reading up on Christo Kutrovsky’s Analysis of the Exadata and Oracle Database Machine announcement from a different point of view.

Also, Alex Gorbachev posted his analysis of the Oracle Database Machine and that’s a worthy read with some new technical details.

You may also be interested in the complete liveblog transcript of the keynote which includes my liveblogging of the rather boring HP advertiseynote before the big show.

I also separated out the Oracle Exadata Storage Server liveblog if you just want to get to the rest of the juicy stuff.

On with the liveblogging!

Second product announcement: The Oracle Database Machine (in partnership with HP).

Specs slides.

8 64-bit servers, 14 exadata storage servers, tons of ram.

Larry: “It will hold really a lot of songs”.

Three year development program.
Custormers:
Amazon, Yahoo, Countrywide, NPD, Quelle

M-Tel: a Bulgarian company, 10-72 times speedup. The worst speedup was 10x.

Alex G: “It’s mainframes!”

Larry: “Next slide”. Martin W: “why doesn’t he have the clicker himself!?”

Christo (a bulgarian) “I can’t believe M-Tel had this and managed to keep it a secret. That’s funny.”

28x P-Series competitive advantage at M-Tel. For half of one.

Darrin L: “Yeah but what’s the price difference. How do you license this!?”

Good questions IMHO. Oracle licenses based on server performance. There is a major problem to be solved there.

TPC-H query set. 30x average speed-up.

They have a paper describing it. Will be intersting to read it.

Christo: “This seems to be parallel-query only! It might be only for data warehouses!”

Larry: “With a conventional array, when you add storage, you don’t add data bandwidth. With this solution, every time you add storage server, you are adding not just disk capacity, but two infiniband pipes, two processors, and more cache”.

Makes Sense. Refer to BAHD again. Man I feel a bit smart right now.

Now bashing Teradata. I guess we’re into the advertising section.

Now Larry is saying how similar this is to Netezza. One processor per disk drive. Christo is saying it’s very similar.

Larry: two big differences:
1. our database machine runs oracle, theirs does not.
2. we lose a drive, ours keeps running, theirs does, their queries stop.
Christo “umm no they don’t they have three way mirroring”.

Great quip: “even I studied about B-Tree indexes in School!” LOL LOL

Slide - comparing vs. Netezza. Clearly this is the company they are targeting.
Oracle’s stuff is bigger and faster and has more cores and faster bandwidth.

Christo: “The question is, does that CPU sort.”

HP Oracle … 650,000 vs. 1.500,000 for Teradata system with less spec.

In the Oracle column, you need to pay the 1.7 mm software license.

Software license for Oracle though… 1.7mm. There is something interesting going on here.

“Even if you pay list for Teradata, it’s cheaper for capacity.”

Next slide……..

Will speed up OLTP as well as data warehousing.

Available today, 10-50x faster than current (in small type) oracle data warehouses”

I must have missed something. How is this not costing more than Teradata?

OK Christo explained, the Oracle system had triple the storage for modestly more cost.

Mark Hurd talking again. Can HP please hire somebody exciting? Maybe exciting is just not what they do.

Adds the fact that these storage servers are “completely open, proliant-based servers”.

I wonder how long this will be HP-only.

Next, I think you’ve listened to me for long enough, now take a moment and listen to this advertising interlude.

More to come.

Liveblogging Larry Ellison’s Keynote

By Paul Vallee September 24th, 2008 at 4:32 pm
Posted in Group Blog PostsOracle
Tags:

I’ll be liveblogging the keynote here along with Marc Fielding, Christo Kutrovsky, Darrin Leboeuf and Luke Davies and Martin Wisniewski.

First observation: I can not believe I can not make this video full screen.

Even Pythian’s flash video player can go full screen.

Introduction time. Safra Katz.

Confirming HP is here as part of the announcement.

So presentation from HP is here first. Weird. Maybe this is a megabucks commercial keynote. Introducting Exec VP Anne Livermore.

“Very very very very very important event for HP. 100000 joint customers.”

Show of hands for HP adoption. A bit lame, she seems disappointed with how many hands are raised. :-)

The video stream is super saturated, the video is super skippy. Hopefully I can understand what’s going on well enough to do this liveblog.

So far it’s completely a vendor advertisement keynote.

“By 2010, more than 1/3 of CEOs and CIOs realize their datacenters will not meet their demand”.

This is not that big a deal according to Marc Fielding. Basically, Duh, 2/3 of datacenter CEOs do not need to invest further to make it through two years, 1/3 does. I agree.

Ugh another advertising slide. I hope there’s something meaty here.

One in six HP servers runs Linux. That’s interesting.

Bragging about HP winning vs. EMC.

HP would be the sixth largest software company in the world if you looked at only software revenue (interesting)

“Transforming the data center”

“True 24×7 lights out automated and energy efficient data center”

I am IMing with Alex Gorbachev who is at the keynote floor.

I asked him:

Paul Vallee
5:44
do you think HP is only going to advertise here?
5:41 PM

or are they part of this x

oracloid11g
5:44
I think they are

I think it’s crazy for Larry to allow HP to announce it then. He will come onstage with them?

Oh my god she’s actually playing an advertisement.

“next generation data center”

Christo: “I think we were too optimistic, thinking it was going to some amazingly cool piece of Oracle software. I’m starting to think we will be disappointed and it will be hardware.”

Alex Gorbachev is twittering live: http://twitter.com/alexgorbachev.

Darrin Leboeuf: “Maybe Oracle bought HP’s services arm!!! LOL”

HP spends 2% of revenue on IT. Interesting. Consolidated 60+ datacenters into 6.
BTW among Pythian’s customers that share such data, I am aware of one large company spending 1.4%.
Not as large as HP though. :-) Maybe there is a cost to scale.

Talking about virtualization now. Still completely unrelated to X as far as I can see.

So far this looks like one of those million dollar advertising keynotes.

29 minutes of this and I’m still awake. I’m pretty proud.

LOL silence…. then “can i have the next chart please”

Christo Kutrovsky: “Can we have the big news now?”
Darrin Leboeuf: “Can we have the next slide please??? please? Next?”

Now she’s talking about EDS. I missed the point as to why?

I guess they’re partnering with EDS on services. All-righty then. This is also not keynote-worthy material.

OK I get it thanks to Martin. HP bought EDS. How did I miss this news? :-)

17:56 Shilling for EDS’ outsourcing. Here’s a tip, choose Pythian instead.

“Performance Optimized Datacenters, our PODs”. Clever.

So HP is following Sun’s lead in container-based datacenters. Way to go.

17:57 Alex Gorbachev: “So boring.”

(Christo suggested I timestamp. All times EST.)

HP working on eliminating copper in computers, replacing with laser+optical, as a power-saving play. Interesting.

18:00 “to wrap up…” woo hoo

Everyone here in my office left except Christo and I. That should paint a picture.

Polite applause.

OK, Larry’s sailboat is on the screen.

“Extreme Performance” . It’s showtime.

Lots of sailboat visuals. Larry loves his sailboat. He seriously loves it. We get it.

18:03 “Ladies and gentlemen, Larry Ellison”

Wow now he’s talking about sailing. Incredible.

OK I’m giving up the sarcasm. The commercial is over and I’m getting ready to get excited.

“next slide please” wow this has not been rehearsed much. Probably because of the secrecy.

Looks like this will be about VLDB. Databases are tripling in size every two years.

“Disk systems today simply can not cope with the amount of data that has to be moved off those drives. We have a huge bandwidth problem”.

“You don’t have to have a 200tb db to start experiencing the slowdowns. 1tb is the elbow of the curve [pretty graph]”

Darrin and Marc are back. I guess it’s getting interesting again.

Two possible solutions:
1. reduce data going through to the storage systems.
2. wider, faster pipes, and more of them.

(reminds me of the bahd).

Announcing Oracle’s first ever hardware product.

The exadata programmable storage server.

Building intelligence into the storage server.

Allows us to reduce the amount of data.

Confirming HP is the partner.

Storage server does not pass disk blocks back to the database server, it actually passes query results.

Note: A few startups are doing this sort of thing already. They should be totally freaking out right now.

Slide explaining how query processing works in traditional storage.

Stark contrast to a grid of exadata storage servers, with processing ability local to each and every disk drive.

“We actually pass the query from the database server directly into the storage servers.”

Explaining how this works.

This reminds me hugely of kickfire for mysql, but for Oracle.

Cool, they’re provisioning two infiniband pipes per storage server. Nice. 40gbps.

Marc Fielding: “The problem is still the disk drives.”

Larry: 1gb/s per exadata storage server, you can have dozens working in parallel.

Christo: 1gb/s??? that’s not that good.

Immediately available for Linux., will work with any Oracle database.
Available for x86. Christo: “not 64 bit???????”

By the way, I made a bit with Paul Cunningham that Kickfire would fail because of Moore’s law. I wonder if I should bet against this tech too.

Second product announcement: The Oracle Database Machine (in partnership with HP).

Specs slides.

8 64-bit servers, 14 exadata storage servers, tons of ram.

Larry: “It will hold really a lot of songs”.

Three year development program.
Custormers:
Amazon, Yahoo, Countrywide, NPD, Quelle

M-Tel: a Bulgarian company, 10-72 times speedup. The worst speedup was 10x.

Alex G: “It’s mainframes!”

Larry: “Next slide”. Martin W: “why doesn’t he have the clicker himself!?”

Christo (a bulgarian) “I can’t believe M-Tel had this and managed to keep it a secret. That’s funny.”

28x P-Series competitive advantage at M-Tel. For half of one.

Darrin L: “Yeah but what’s the price difference. How do you license this!?”

Good questions IMHO. Oracle licenses based on server performance. There is a major problem to be solved there.

TPC-H query set. 30x average speed-up.

They have a paper describing it. Will be intersting to read it.

Christo: “This seems to be parallel-query only! It might be only for data warehouses!”

Larry: “With a conventional array, when you add storage, you don’t add data bandwidth. With this solution, every time you add storage server, you are adding not just disk capacity, but two infiniband pipes, two processors, and more cache”.

Makes Sense. Refer to BAHD again. Man I feel a bit smart right now.

Now bashing Teradata. I guess we’re into the advertising section.

Now Larry is saying how similar this is to Netezza. One processor per disk drive. Christo is saying it’s very similar.

Larry: two big differences:
1. our database machine runs oracle, theirs does not.
2. we lose a drive, ours keeps running, theirs does, their queries stop.
Christo “umm no they don’t they have three way mirroring”.

Great quip: “even I studied about B-Tree indexes in School!” LOL LOL

Slide - comparing vs. Netezza. Clearly this is the company they are targeting.
Oracle’s stuff is bigger and faster and has more cores and faster bandwidth.

Christo: “The question is, does that CPU sort.”

HP Oracle … 650,000 vs. 1.500,000 for Teradata system with less spec.

In the Oracle column, you need to pay the 1.7 mm software license.

Software license for Oracle though… 1.7mm. There is something interesting going on here.

“Even if you pay list for Teradata, it’s cheaper for capacity.”

Next slide……..

Will speed up OLTP as well as data warehousing.

Available today, 10-50x faster than current (in small type) oracle data warehouses”

I must have missed something. How is this not costing more than Teradata?

OK Christo explained, the Oracle system had triple the storage for modestly more cost.

Mark Hurd talking again. Can HP please hire somebody exciting? Maybe exciting is just not what they do.

Adds the fact that these storage servers are “completely open, proliant-based servers”.

I wonder how long this will be HP-only.

Next, I think you’ve listened to me for long enough, now take a moment and listen to this advertising interlude.

My comments: This is definitely a premium play. It will be interesting to follow the developments. It is not at all what I thought it would be, but it is fascinating.

I bet against Kickfire with Paul McCullagh because of Moore’s Law limiting the long-term value of the speed-up. Maybe I was wrong, Larry is betting a lot more money that I am.

Christo: Let me sum it up: This is parallel query being pushed down to the disk.

Christo: I have two words for you: Object Checkpoint.
Basically before you run a parallel query on any object, you have to checkpoint it. This is already there in 10g and is an enabling technology.

Darrin: When you start up an instance, does it start up an instance on the disk server?
Christo: Doesn’t need it, just needs a filter server.

Christo: The question is, does it sort???????? This is extremely important.

Christo: My first guess was a SAN. But I never imagined Oracle would start having a hardware play. (Or two).

Christo: They got tired of people not buying bandwidth, so they’re forcing people to buy it now. Smart.

Looks like Larry will be talking about X with HP

By Paul Vallee September 24th, 2008 at 3:39 pm
Posted in Oracle
Tags:

I’m sorry if everyone else knew this already, but I just noticed the following from the Live Keynote page (click on Larry’s keynote to see it):

Larry with HP about X

So Larry will be joined on stage by bigwigs from HP.

You will remember that HP bought Polyserve, Kevin Closson’s clustered file systems company, a little while back.

No coincidence, I think, but as we know from multiple sources, Kevin is now at Oracle as an Architect on this new technology that Larry is announcing.

At first I thought the HP keynote was one of those big-pay vendor closers. But now, with it being one session in the Live View, and what we’re starting to learn about the likely nature (storage) of the innovation, the following snippet can easily be re-interpreted (interesting bit in my bold):

Transforming Business and Technology Today and Tomorrow
Innovation is the lifeblood of information technology, but businesses are far more selective today in the kinds of technology they will buy and deploy to ensure they remain competitive. They’re looking for practical innovation that will optimize business results such as lowering IT costs, reducing risk and improving growth and profitability. Come learn about HP’s customer-focused innovation, including investments in research, product development and advanced services that have increased energy efficiency, provided new approaches to datacenter transformation, and also given us new advances in internet technology that are shaping the future of enterprise IT. Also highlighted: the joint innovation that HP and Oracle are delivering to their customers around the world.

Ann Livermore | Executive Vice President, Technology Solutions Group, HP | Biography [+/-]
Mark Hurd | Chairman of the Board and CEO, HP

The pieces are coming together. Stay tuned as I will be writing further about this as the keynote begins.

Much Oracle Ado

By Keith Murphy September 23rd, 2008 at 10:46 am
Posted in MySQLNon-Tech ArticlesOracle
Tags:

If you track the database world outside of MySQL, you know that Oracle is having a conference this week. It’s called Oracle Open World. Drips with irony doesn’t it? But this post isn’t about Oracle being open or otherwise.

This post is about the announcement being made Wednesday. It seems Oracle has a surprise. A pretty well kept surprise. It’s such a big deal that Larry Ellison himself is making the announcement.

Some people, including some of my colleagues at Pythian, are speculating that this is going to be an announcement about a share-nothing clustering solution.

In the first quarter of 2007, I interviewed with a company in Atlanta, seeking my first full-time job as a MySQL database administrator. They were an online company building a social-networking website with a virtual world interface (kind of like Second Life, from what I understood). They were using an (at the time) fairly unstable version of MySQL 5.1 only because it offered clustering with the ability to store data on disk while keeping the indexes in memory. Previously, in version 5.0, everything had to be stored in-memory. Much has improved with MySQL clustering since that time.

While I don’t know for certain that Larry is going to announce in-memory clustering, I kind of hope that is what it’s all about, because it would demonstrate this: Oracle is walking a trail blazed by MySQL.

OOW Video: Gavin Newsome and Ed Begley, Jr. Talk About Being Green

By Sheeri Cabral September 22nd, 2008 at 7:57 pm
Posted in Non-Tech ArticlesOracle
Tags:

Gavin Newsome, mayor of San Francisco, and Ed Begley, Jr. talk about being green, and commend Oracle and Oracle OpenWorld for being green.

Watch the video online or download the 19 Mb flash video file

Watch the video online or download the 18 Mb flash video file

(more…)

Oracle’s Secret New Feature: Educated Guesses

By Christo Kutrovsky September 22nd, 2008 at 3:52 pm
Posted in MySQLNon-Tech ArticlesOracle
Tags:

Larry Ellison is announcing a major new feature this Wednesday at Open World. For the first time in a while, his keynote is dedicated to the “database” as opposed to the usual high level ERP/Apps/Fusion. Even the title of his keynote is catchy — “Extreme Performance”.

Oracle has been keeping the new feature a secret. Even the 11gR2 beta program had very few participants to prevent information leaking out. It’s, “Something’s coming, but I am not telling what.”

Okay, it worked on me, I’m excited about it. Let’s think what it could be. What single database feature is so major, that Larry himself will announce it during OpenWorld?

What do we know so far?

  • Starting with the obvious, Larry’s keynote is “Extreme Performance”, so it’s related to performance.
  • We know Kevin Closson has worked on it - he had a blog entry saying “I am working on something big” that got pulled off the web. (Here’s Google’s cache.)

Given these two point, let’s further think about it. What do we know about Kevin?

  • He worked for PolyServe — a company whose main product is a cluster file system.
  • He worked for Sequent on NUMA systems, which in today’s world is pretty close to cluster software with a very fast, low latency interconnect.
  • He is an expert in storage systems and disk performance.
  • He joined Oracle recently, possibly to work on this secret project.
  • He must be really excited about it, to post anything on his blog under radio silence.

I think it’s something related to storage, something new and revolutionary about storage. But what?

We already know, from leaks on certain websites, that ASM will become a cluster filesystem which will allow storing OCR files, as well as user files, on the ASM disks.

But is this big enough? It’s definitely significant. Now you get a “free” reliable, cluster file system with Oracle. I don’t think it’s big enough though. Oracle already had OCFS and OCFS2. So it’s not something new to release a filesystem. And even if ASM becomes a true filesystem, that would not provide such a significant performance boost to warrant a keynote called “Extreme Performance”. An ASM filesystem would be a major manageability feature, not so much a performance feature.

That being ruled out, what could it be?

Recently, when setting up a new 11g database on a server with 128gb of RAM, I was setting up hugepages as usual, and thinking about how big my cache would be. It struck me that the cache will be bigger than the database for quite a while. Why do we even need the SAN/Datafiles?!

Then it hit me.

(more…)

OOW Video: Mary Matalin and James Carville Keynote

By Sheeri Cabral September 22nd, 2008 at 11:20 am
Posted in Group Blog PostsNon-Tech ArticlesOraclePythian Goodies
Tags:

To start off the conference, the first keynote at Oracle OpenWorld took a break from technology and veered into the world of politics. The official conference description says:

Washington’s best-loved political couple Mary Matalin and James Carville entertain the crowd with a bitingly humorous look at the world of politics.

Indeed, there was humor, and politics. For a light-hearted yet factual look at US politics, watch the video by streaming directly in your browser or download the 176 Mb Flash video file.

(more…)

Log Buffer #115: A Carnival of the Vanities for DBAs

By David Edwards September 19th, 2008 at 11:45 am
Posted in Log BufferMySQLNon-Tech ArticlesOraclePostgreSQLSQL Server
Tags:

Welcome to the 115th edition of Log Buffer, the weekly review of database blogs.

I must thank Paul for taking over at last minute for LB#114 last week, when, as he put it, “ . . . a killer combo of painkillers and the pain that the painkillers can’t kill . . . ” reduced to me a less-than Log-Buffer-capable state. Or to be more precise, to a writhing, benighted gargoyle of misery. (Too colorful?)

Anyway, the good news is that I’m better. Not all better, mind you. Between the tooth thing and my spending all my working time on a special project, there was nothing left for poor old Log Buffer. So, I face the choice: throw it open to you, LB’s loyal readers for your contributions; or adopt Paul’s approach1 from last week, and use the nifty AideRSS.

I’m going to bet on our readers. Let’s hear from you with your picks for best database blogs of the week gone by. I promise you a real, proper Log Buffer next week, from someone. If not me, well, Nick Westerlund still wants his go, and Ward Pond is back looking for a slot.

Until then, wish me luck with my angry tooth.

1. The truth is that I was briefly worried about having my job taken away by software. My concerns were allayed, at least partially, when I saw that the original software-built list of database blogs also included an item from “manscaping.com”, which I’m fairly sure had little or nothing to do with database administration.

GNU basename in PL/SQL

By Don Seiler September 17th, 2008 at 5:03 pm
Posted in Oracle
Tags:

In the process of scripting a database migration, I was in need of something akin to the GNU basename utility that I know and love on Linux. basename is most famous for taking a full file path string and stripping away the leading path component, returning just the name of the file. This can be emulated in PL/SQL with calls to SUBSTR and INSTR, like this:

substr(dirname,instr(dirname,'/',-1)+1)

(Thanks to Ian Cary, who shared this logic on oracle-l)

As you can see, this simply finds the last occurence of /, which is our directory separator on *nix and Solaris operating systems. On Windows, it would be \. It then returns a substring beginning one character after that last separator until the end of the string. Voila, a basic basename routine!

Upon reading the basename man page again, I found that basename also takes an optional parameter, a suffix string. If this suffix string is provided, basename will also truncate that string from the end. For example:

(more…)