I'm a member of
Quicksearch
Codenews Search
Disclaimer
The individual owning this blog works at Sun Microsystems GmbH in Germany, a subsidiary of Oracle. The opinions expressed here are his own, are not necessarily reviewed in advance by anyone but the individual author, and neither Oracle nor any other party necessarily agrees with them.
Navigation
Categories
About this blog- Apple
- iPhone
- Aviation
- Bahn
- Begriffe des Lebens
- Blogosphere
- Braindump
- Business Travel
- CANBOX
- CeBIT
- CEC
- Computing
- del.icio.us
- Fundsache
- General
- HHOSUG
- i hate sundays
- Kochen
- Mindfuck
- Movies
- Music
- On Benchmarketing
- Persoenliches Pech
- Photographie
- policy of ...
- Privacy
- Recommended Blogs
- Security
- Solaris
- Strange Days
- Sun/Oracle
- suncec2007
- suncec2008
- Technology
- The IT Business
- Toolbox
- Travel
- Work
Shared sZIL without an external storage box
Thursday, December 3. 2009
Trackbacks
Trackback specific URI for this entry
No Trackbacks
Comments
Display comments as
(Linear | Threaded)
I thought AVS is buggy? Anyways,
Quick question: Is Anago based on Jasper Forest? If it is, it is going to be pretty exciting.
Quick question: Is Anago based on Jasper Forest? If it is, it is going to be pretty exciting.
1. Version 1.0 of the tools included in AVS gave a bad reputation. Not so well informed admins mostly still know about it only from anecdotal comments based on experiences of this time. We are at 4.x is really stable and it's used quite frequently. The toolset does what it should do ...
2. Still selling Nehalems wherever you write? I won't discuss unannounced products with you.
2. Still selling Nehalems wherever you write? I won't discuss unannounced products with you.
I don't care if you believe me or not, I am not affiliated with Intel. I know that, Intel knows that. That's the end of the story.
I am inquiring about Anago because HP announced Jasper Forest based arrays on youtube the other day.
http://www.youtube.com/watch?v=uBOMQ5W5jl8
Joerg, if anything, you should look into your biased tones. You obviously hate Intel, but since you are German, I don't blame you.
Most of the time, your critics do have a point. Just the other day, you told one of your critics to "fuck off"
http://www.c0t0d0s0.org/archives/6114-Im-starting-to-get-really-angry-.....html
That wasn't very nice or professional. Especially that critic had the same opinion as me, that the F5100 wasn't designed well enough to boost performance worthy of its price tag.
You can't really say "fuck off" to anyone who disagree with you.
BTW, "I won't discuss unannounced products".....that's what I expected from you. We will just wait and see the hardware specification in a few months. Jasper Forest isn't exactly NDA material. HP announced it. If Sun is going to adopt that, you might as well announce it too to draw more interest.
I am inquiring about Anago because HP announced Jasper Forest based arrays on youtube the other day.
http://www.youtube.com/watch?v=uBOMQ5W5jl8
Joerg, if anything, you should look into your biased tones. You obviously hate Intel, but since you are German, I don't blame you.
Most of the time, your critics do have a point. Just the other day, you told one of your critics to "fuck off"http://www.c0t0d0s0.org/archives/6114-Im-starting-to-get-really-angry-.....html
That wasn't very nice or professional. Especially that critic had the same opinion as me, that the F5100 wasn't designed well enough to boost performance worthy of its price tag.
You can't really say "fuck off" to anyone who disagree with you.
BTW, "I won't discuss unannounced products".....that's what I expected from you. We will just wait and see the hardware specification in a few months. Jasper Forest isn't exactly NDA material. HP announced it. If Sun is going to adopt that, you might as well announce it too to draw more interest.
Well ... it's just the way, that an affiliation with Intel would explain your behaviour 
The "critic" doesn't got his "Fuck Off" for his thoughts, albeit i wrote a dozen times about the error in his reasoning, he got the "Fuck off" primarily for the accusation of "making up numbers" in Giris blog and secondarily for "faking an discussion". Just to get the impression correct. I think that was a well deserved "Fuck off".
I don't hate Intel. From my point of view, we could even use Intel Atoms, MIPS derivates or whatever you can think of in the 7xxx line. It's about the best decision in the light of a multitude of requirements. You tend to look at the wrong end of technology development: The processor decision is one of the later ones, not the first one. As i wrote before: When Nehalem EX is up to all the requirements (and only few of them are technological ones) then i think we will see an 7xxx based variant. If Rainbow Falls is more up to the task, then i think we will see RF in the boxes. That's the nice thing of using Opensolaris (or any other operating system available on more than one architecture) you are relatively independent of the CPU.
BTW: You could easily answer you question, especially as you tried to sell Nehalem EX before Anago is said to be a highend system in the 7xxx line. Jasper Forest is limited to 2 sockets. That doesn't sound as the correct choice for a high end system.
Jasper Forest isn't NDA material, but Anago is, despite some presentations had leaked. I will not talk about it. Period. And by the way, most customers don't care what proc is in the system. They care about features of the storage system for example.
PS: I have the strange urge to buy a song from the Pet Shop Boys at iTunes - http://bit.ly/8kB6CE
The "critic" doesn't got his "Fuck Off" for his thoughts, albeit i wrote a dozen times about the error in his reasoning, he got the "Fuck off" primarily for the accusation of "making up numbers" in Giris blog and secondarily for "faking an discussion". Just to get the impression correct. I think that was a well deserved "Fuck off".
I don't hate Intel. From my point of view, we could even use Intel Atoms, MIPS derivates or whatever you can think of in the 7xxx line. It's about the best decision in the light of a multitude of requirements. You tend to look at the wrong end of technology development: The processor decision is one of the later ones, not the first one. As i wrote before: When Nehalem EX is up to all the requirements (and only few of them are technological ones) then i think we will see an 7xxx based variant. If Rainbow Falls is more up to the task, then i think we will see RF in the boxes. That's the nice thing of using Opensolaris (or any other operating system available on more than one architecture) you are relatively independent of the CPU.
BTW: You could easily answer you question, especially as you tried to sell Nehalem EX before
Anago is said to be a highend system in the 7xxx line. Jasper Forest is limited to 2 sockets. That doesn't sound as the correct choice for a high end system.Jasper Forest isn't NDA material, but Anago is, despite some presentations had leaked. I will not talk about it. Period. And by the way, most customers don't care what proc is in the system. They care about features of the storage system for example.
PS: I have the strange urge to buy a song from the Pet Shop Boys at iTunes
- http://bit.ly/8kB6CE
"Most customers don't care what procs is in the system" -Joerg
That statement is only true if Sun expects its customers to be morons. After all, Sun is a computer maker. If Sun doesn't think that its customers care about what processor they use, isn't that precisely the reason why Sparc had been losing market share all these years?
The reason why I am interested in Jasper Forest is the integrated RAID acceleration, and non-transparent bridging linking two systems together for High Availability. (I don't know yet if that's single system image)
It is flawed to think that the 2Socket Jasper Forest is not high end. It will benchmark very close to your highest end 4 Socket Istanbuls 7410s, and will be faster than your last year's Opteron 8356 based 7410s.
Anyways, even though you refused to talk about unannounced products with me, technically you told me everything I need to know about Anago. I got to go. Later.
That statement is only true if Sun expects its customers to be morons. After all, Sun is a computer maker. If Sun doesn't think that its customers care about what processor they use, isn't that precisely the reason why Sparc had been losing market share all these years?
The reason why I am interested in Jasper Forest is the integrated RAID acceleration, and non-transparent bridging linking two systems together for High Availability. (I don't know yet if that's single system image)
It is flawed to think that the 2Socket Jasper Forest is not high end. It will benchmark very close to your highest end 4 Socket Istanbuls 7410s, and will be faster than your last year's Opteron 8356 based 7410s.
Anyways, even though you refused to talk about unannounced products with me, technically you told me everything I need to know about Anago. I got to go. Later.
Sorry ... it looks like you don't understand this part the IT business. The 7000er line is a storage appliance. Do you care what CPU is in your washing machine, in your switch, in your car?
As long as the system delivers the data in a speed expected by the customers, it isn't important what CPU is in the system. If it's MIPS, if it's Power, if it's SPARC, if it's AMD, if it's Intel. Customers out of the realm of the home enthusiast market don't care for the proc, as long it fullfils their requirements as defined by the business stakeholders.
As i told you the the logic goes: Hardware follows OS, OS follows Middleware , Middleware follows Application, Application follows business demands ... not the other way round ....
I will tell you an experience from the field: Telling a decisionmaker "It's a Nehalem EPXKLTE in it with superduper 3 GHz and a three channel memory interface in it" and she or he will just yawn. Telling him "We are able to cope with the load imposed by your SAP, your Exchange and your Windows Fileserving, we fit in your operational guidelines, and you save 30% annually" you get an interested look. Sorry, as much as i like to talk about technology as well but that is the harsh reality of enterprise IT.
At the end this has much to do with responsibility: When you go to a vendor and tell him "I need 20 X4275 and two 6140" then you, the customer has the responsibility for the correctness of the sizing, if a customer tells a vendor "I need a new architecture for my SAP. I need the same processing time than today plus 40%. Do whatever you want." it's the responsibility of the vendor and the customers has a "go-out-of-jail"-card when the system isn't up to the task.
That's not about morons, it's about business. The only people interested in the exact CPU are the HPC people, as their mostly have codes that are vastly tuned to a certain architecture and as they know exactly the performance characteristics of their code in conjunction with CPUs, numbers of cores, cache sizes. Many other people just look at the software and tell me "This is my software, this is the version" and then i start to look in support matrices.
I can live with this disclosure ... that one was sooo obvious
I won't start any discussion about Nehalem, Nehalem-EX, Opteron with you as it doesn't look that you didn't get the basic point of the appliance business: The hardware doesn't matter. Provided services, fullfilment of requirements and price matters.
As long as the system delivers the data in a speed expected by the customers, it isn't important what CPU is in the system. If it's MIPS, if it's Power, if it's SPARC, if it's AMD, if it's Intel. Customers out of the realm of the home enthusiast market don't care for the proc, as long it fullfils their requirements as defined by the business stakeholders.
As i told you the the logic goes: Hardware follows OS, OS follows Middleware , Middleware follows Application, Application follows business demands ... not the other way round ....
I will tell you an experience from the field: Telling a decisionmaker "It's a Nehalem EPXKLTE in it with superduper 3 GHz and a three channel memory interface in it" and she or he will just yawn. Telling him "We are able to cope with the load imposed by your SAP, your Exchange and your Windows Fileserving, we fit in your operational guidelines, and you save 30% annually" you get an interested look. Sorry, as much as i like to talk about technology as well but that is the harsh reality of enterprise IT.
At the end this has much to do with responsibility: When you go to a vendor and tell him "I need 20 X4275 and two 6140" then you, the customer has the responsibility for the correctness of the sizing, if a customer tells a vendor "I need a new architecture for my SAP. I need the same processing time than today plus 40%. Do whatever you want." it's the responsibility of the vendor and the customers has a "go-out-of-jail"-card when the system isn't up to the task.
That's not about morons, it's about business. The only people interested in the exact CPU are the HPC people, as their mostly have codes that are vastly tuned to a certain architecture and as they know exactly the performance characteristics of their code in conjunction with CPUs, numbers of cores, cache sizes. Many other people just look at the software and tell me "This is my software, this is the version" and then i start to look in support matrices.
I can live with this disclosure ... that one was sooo obvious
I won't start any discussion about Nehalem, Nehalem-EX, Opteron with you as it doesn't look that you didn't get the basic point of the appliance business: The hardware doesn't matter. Provided services, fullfilment of requirements and price matters.
Sorry, just got back. I enjoy arguing with you. Unfortunately, due to the fact that you work for Sun, your view is inherently biased when we talk about Sun products.
"Sorry ... it looks like you don't understand this part the IT business. The 7000er line is a storage appliance. Do you care what CPU is in your washing machine, in your switch, in your car?"
I do understand this fudging part of business perfectly. Sun is trying to "appliantize", if that's a word, a Sun Fire Storage server to fudge more margins out of it. It is just another computer with some software in the end. I don't care what CPU is in my washing machine. But I do care about the CPU in the switch, because that will ultimately determine PPS. As to the CPU in the car, you are making the wrong comparison. The CPU in a server is like a engine in a car. Of course people care about the maker of the engine right? I am not a car guy myself, but I would think people make discriminatory choices when looking at the horsepower of the engine and maker of the engine(HEMI? I know I am not that technical when it comes to cars)
"As long as the system delivers the data in a speed expected by the customers, it isn't important what CPU is in the system. If it's MIPS, if it's Power, if it's SPARC, if it's AMD, if it's Intel. Customers out of the realm of the home enthusiast market don't care for the proc, as long it fullfils their requirements as defined by the business stakeholders."
This argument is often used by non-technical people to deceive morons. When designing a storage "appliance" or a "car", if there are two choices available to choose for the "engine" of the appliance, one is faster and more than 20% cheaper and eats less "gas"(Nehalem-EP L5520 vs Istanbul 2431 particularly interesting), if you are a responsible car designer looking at this specs, which one would you choose? That depends on the principle of the designer. If the designer has some sort of bias, paid or, or unpaid for, to choose the technically weaker products because he thinks that none of his customers would know or care about his chocies, then it would only attract the kind of customers who don't know anything or are foolish enough to believe that the Open Storage 7410 is an appliance, thus worth 10x times more than the X4440 server underneath.
"That's not about morons, it's about business."
Absolutely. If Sun believes that business is about making retarded and biased product choices and hoping that customers lack the intelligence to figure it out, then, it is the perfect explanation why Sparc is losing market share and Sun is losing x86 market share and the fact that Sun is losing gobs of money. Even though its software products are tremendous(ZFS, Solaris).
You want to talk about hardware? You know why the F5100 is slow? Because the designer was stupid enough to use SAS 1 Expanders and SAS 1 HBAs to drive 80 FMODs which is another form factor change. Even if you look at the FMODs, it is 8 channels SLC with 2 channels sacrificed, and driven by a Marvell controller(I would have gone with Indilinx). In the end you got 80 FMODs trying to squeeze through the SAS1 pipe, jacking up the latency to 400us+, that's why it is slow, and requires 16 HBAs to drive it to the "advertised" 1.6Million IOPs. On the other hand, if you are really determined to use 16 HBAs to get to the 1.6 Million IOPs, you could have just used 16 F-20s with the FMODs directly attached to the HBA without going through the SAS expander. That solution would have given you 64 FMODs directly attached without the expander latency. If you go down that route, even fusion IO is faster.
Going forward, I think Witte is a great product. If Sun allows people to fill the Witte X25-E G2 50GB SSDs and use it as L2ARC for its Anago platform, at least people would consider that proposition. Math would easily predict that 3x Witte with X25-Es would beat the F5100 at 20% of the cost of the F5100.
The rest I leave it to the Fishworks engineers to think about.
"Sorry ... it looks like you don't understand this part the IT business. The 7000er line is a storage appliance. Do you care what CPU is in your washing machine, in your switch, in your car?"
I do understand this fudging part of business perfectly. Sun is trying to "appliantize", if that's a word, a Sun Fire Storage server to fudge more margins out of it. It is just another computer with some software in the end. I don't care what CPU is in my washing machine. But I do care about the CPU in the switch, because that will ultimately determine PPS. As to the CPU in the car, you are making the wrong comparison. The CPU in a server is like a engine in a car. Of course people care about the maker of the engine right? I am not a car guy myself, but I would think people make discriminatory choices when looking at the horsepower of the engine and maker of the engine(HEMI? I know I am not that technical when it comes to cars)
"As long as the system delivers the data in a speed expected by the customers, it isn't important what CPU is in the system. If it's MIPS, if it's Power, if it's SPARC, if it's AMD, if it's Intel. Customers out of the realm of the home enthusiast market don't care for the proc, as long it fullfils their requirements as defined by the business stakeholders."
This argument is often used by non-technical people to deceive morons. When designing a storage "appliance" or a "car", if there are two choices available to choose for the "engine" of the appliance, one is faster and more than 20% cheaper and eats less "gas"(Nehalem-EP L5520 vs Istanbul 2431 particularly interesting), if you are a responsible car designer looking at this specs, which one would you choose? That depends on the principle of the designer. If the designer has some sort of bias, paid or, or unpaid for, to choose the technically weaker products because he thinks that none of his customers would know or care about his chocies, then it would only attract the kind of customers who don't know anything or are foolish enough to believe that the Open Storage 7410 is an appliance, thus worth 10x times more than the X4440 server underneath.
"That's not about morons, it's about business."
Absolutely. If Sun believes that business is about making retarded and biased product choices and hoping that customers lack the intelligence to figure it out, then, it is the perfect explanation why Sparc is losing market share and Sun is losing x86 market share and the fact that Sun is losing gobs of money. Even though its software products are tremendous(ZFS, Solaris).
You want to talk about hardware? You know why the F5100 is slow? Because the designer was stupid enough to use SAS 1 Expanders and SAS 1 HBAs to drive 80 FMODs which is another form factor change. Even if you look at the FMODs, it is 8 channels SLC with 2 channels sacrificed, and driven by a Marvell controller(I would have gone with Indilinx). In the end you got 80 FMODs trying to squeeze through the SAS1 pipe, jacking up the latency to 400us+, that's why it is slow, and requires 16 HBAs to drive it to the "advertised" 1.6Million IOPs. On the other hand, if you are really determined to use 16 HBAs to get to the 1.6 Million IOPs, you could have just used 16 F-20s with the FMODs directly attached to the HBA without going through the SAS expander. That solution would have given you 64 FMODs directly attached without the expander latency. If you go down that route, even fusion IO is faster.
Going forward, I think Witte is a great product. If Sun allows people to fill the Witte X25-E G2 50GB SSDs and use it as L2ARC for its Anago platform, at least people would consider that proposition. Math would easily predict that 3x Witte with X25-Es would beat the F5100 at 20% of the cost of the F5100.
The rest I leave it to the Fishworks engineers to think about.
Links in this article
The LKSF book
The book with the consolidated Less known Solaris Tutorials is available for download here
Twitterfeeds
twitter.com/c0t0d0s0
just blogged: Reengining: I've learned in the last few weeks that a simple design decision can make your life much... http://bit.ly/d9luy8
twitter.com/codenews
6935782 need to manually increment build number one more time http://bit.ly/aMqEbX
twitter.com/SunPatches
128365-04 - Sun Crypto Accelerator 6000 1.1: Driver Patch. Available for SPARC since Mar/19/10. http://bit.ly/agl9Nw
twitter.com/SolPatchesX86
118192-04 - SunOS 5.9_x86: gtar patch. Available since Mar/19/10. http://bit.ly/cbnoJ7
twitter.com/SolPatchesSPARC
118191-04 - SunOS 5.9: gtar patch. Available since Mar/19/10. http://bit.ly/cb2Drj
just blogged: Reengining: I've learned in the last few weeks that a simple design decision can make your life much... http://bit.ly/d9luy8
twitter.com/codenews
6935782 need to manually increment build number one more time http://bit.ly/aMqEbX
twitter.com/SunPatches
128365-04 - Sun Crypto Accelerator 6000 1.1: Driver Patch. Available for SPARC since Mar/19/10. http://bit.ly/agl9Nw
twitter.com/SolPatchesX86
118192-04 - SunOS 5.9_x86: gtar patch. Available since Mar/19/10. http://bit.ly/cbnoJ7
twitter.com/SolPatchesSPARC
118191-04 - SunOS 5.9: gtar patch. Available since Mar/19/10. http://bit.ly/cb2Drj
Web 2.0
Contact
Networking
open.bc
LinkedIn
Facebook
My photos
- mail: joerg@c0t0d0s0.org
- icq: 144000169
- Me, myself and I
- openBC
- My del.icio.us
- c0t0d0s0@twitter
- c0t0d0s0@last.fm
- wiki@c0t0d0s0
- Amazon Wishlist
Networking
open.bc
My photos
Syndication
Tagged articles
AMD Apple avs Bahn Blogging Blogosphere braindump Business Travel CeBIT cec cec2006 CMT del.icio.us deutsch dtrace fliegen Fundsache General Hamburg IBM i hate sundays Intel iscsi jumpstart Links Linux lksf Mindfuck Movies Music Musik Niagara Opensolaris Opteron Photographie policy of ... Politik Security Solaris storage Sun suncec2007 sunw t1 The IT Business Ultrasparc ultrasparc t1 Wirtschaft Work ZFS
Comments
Sat, 20.03.2010 08:55
Yes. And I just don't like the
way they're killing all of Su
n brands.
They could just buy
, help, let live, contro [...]
Sat, 20.03.2010 08:49
Well, I don't think many peopl
e were using Solaris at home b
efore Oracle acquisition too,
I see home servers more [...]
about Who are you?
Sat, 20.03.2010 02:15
Ich bin im Rahmen der Diskussi
on um das Zugangserschwerungsg
esetz auf dein Blog gestoßen.
Als Linux-Begeisterter d [...]
Sat, 20.03.2010 00:32
The article doesn't explain wh
y the adquisition of Sun is go
ing to be a sucessfull. It onl
y says that we all know: [...]
Fri, 19.03.2010 20:58
Well, I am being paid to take
care of Solaris 10 systems and
my company will continue to u
se it. But the relativel [...]
Buttons
This work is licensed under a Creative Commons Attribution-Noncommercial-Share Alike 3.0 Germany License
