Slashdot Mirror
HP Launches Moonshot
New submitter linatux writes "HP has announced their 'Moonshot 1500 server' — up to 1,800 servers per 47U rack are supported. The tech certainly seems to be an advance on what is currently available — will it be enough to revive HP's server fleet?"
From Phoronix: "Moonshot began with Calxeda-based ARM SoCs, but in the end HP settled for Intel Atom processors. Released today were HP's Moonshot system based on the Intel Atom S1200. Hewlett-Packard claims that their Moonshot System uses 89% less energy, 80% less space, 77% less cost, and 97% less complexity than traditional servers."
Low power and massive amounts of parallel cores is alright, but does it compute? How do these low power servers benchmark against EC2 or equivalent? This article didn't talk benchmarks. Maybe you get all these gains in consumed power, cost, space etc... because it is 90% less powerful than competitors.
Build a man a fire and you warm him for a day. Set a man on fire and you warm him for the rest of his life.
And we all know that datacenter servers are used only for running the front-end of 3D games. Oh no, this product must be DOOMED!
Atom processors are notoriously slow. You can't play 3d video games on them.
Yes, you can. :-)
I managed to play Orbiter on a reasonable resolution (1280x1024x16) and got an acceptable (barely, I admit) framerate on my Atom 330 box. That it's my Media Center and torrent server, by the way.
Granted, the Game of the Year will not run on this setup. :-)
Lisias@Earth.SolarSystem.OrionArm.MilkyWay.Local.Virgo.Universe.org
HP tried this with Transmeta a while back, and produced blades that completely sucked - WAY too slow. Individual machines on blades are dead, unless you need HPC type power, and Atom ain't that. If you need to squeeze 1800 limp servers into a rack, VMWare and its children are already there.
Sorry HP, you suck. Go back to making shitty printers, and then get out of the way. Hopefully your corpse will provide the fertilizer for some new market leader to grow from.
I want to delete my account but Slashdot doesn't allow it.
Whenever slashdot asks "Will it be enough?" what do we say everybody? NO! We say N-O. No.
HP has been attracting fail like it's a government project with unlimited funding and no congressional oversight. I mean seriously, we may be breaking into new physics here with the strong attractive force that all things HP have to all things Fail. And no technology is going to fix that, because the ultimate source of the bogon radiation is (wait for it) HP senior management. They'll figure out a way to screw this up, trust me. They could have just discovered the Holy Grail and they'd still somehow figure a way to botch it so instead of getting eternal youth, we're stuck with an endless series of ever more-expensive drinking glasses that can only hold certain types of beverages and occasionally explode for no reason.
#fuckbeta #iamslashdot #dicemustdie
The Money Shot?
"To those who are overly cautious, everything is impossible. "
From the HP Site: "The HP Moonshot 1500 Chassis has 45 hot-pluggable servers installed and fits into 4.3U. The density comes in part from the low-energy, efficient processors. The innovative chassis design supports 45 servers, 2 network switches, and supporting components."
Each pluggable unit support 1x 2GHZ intel atom S1200 series cpus (2x core, 4x thread), up to 1 dimm @ 8gb, and one SFF sata drive. That gives you 90 cores/180 threads, 360GB's in 4.3u.
For comparison a 6RU cisco UCS chasis can put down up to 160 Cores / 320 threads, 4TB of memory. Those are high performance Xeon cores. Not sure on the $$$ per compute/memory between the two.
The really big question is are there enough use cases for that many "thin" servers. At 2 cores and 8GB of ram you are very thing by modern standards and there is 0 opportunity for vertical growth.
Just how much time your multi threaded server waiting for storage or cache miss? Hint, a lot!
Tomorrow is another day...
Would somebody please find the marketers/editors that wrote this and shoot them? THXBYE
Hi. I'm part of the engineering team tasked with tracking down and eliminating people upon request, who have managed to slight someone else on the internet. We've logged your request and will get to it as quickly as possible. However, due to our limited budget and the unexpected popularity of our service, the high volume of requests will delay our response time. We currently estimate that we'll be able to service your request on October 27th, 2238, at 8:00 pm.
Also, our records indicate that you have an appointment with us on June 27th, 2027, at 3:35pm. Please be prompt. The internet as you know it depends on it.
#fuckbeta #iamslashdot #dicemustdie
So...former HP customer, or former employee?
I'm guessing you haven't actually used HP servers or compared them to the competition. In my experience they completely kick Dell's butt, and give IBM a real run for their money, at much lower cost. I evaluated a ProLiant Gen8 and the manageability features were pretty impressive. The thing can update it's firmware and send SNMP traps, etc, from bare metal, without an OS.
Granted, HP had some crappy CEOs, and on the low-end consumer stuff they race to the bottom with everyone else, but their servers are serious and arguably industry-leading. They also sell more PCs than anyone anywhere, unless you start counting every iPod touch as a "computer."
Lots of datacenter servers are mostly used for holding a connection open while it waits for the database to spit back a resultset. Hense products like these.
Someone at Phoronix really needs to learn basic math. The Chassis is 4.3U and hold 45 of these Moonshot servers, so a 47U rack could fit 10 chassis' for a total of 450 servers.
A game has objectives and is competitive, anything else is just play
Nope. http://cache.gizmodo.com/assets/images/8/2008/04/Melies_TripMoon_largest.jpg
Nazi is a proper noun and thus must be capitalized.
And yet you don't know how to spell "capitalised". Bloody colonials...
I dunno.
The HP P4000s SANs are pretty nice when compared with comparable equipment.
Of course, they got them by buying LeftHand.
But yeah, long gone are the days of the solid Laserjet 4250 days with millions of prints that made them worth refurbishing.
THL phish sticks
But in all seriousness, this is a great idea for crowd sourcing.
Acquire startup funding and open a website for sponsors and volunteers. You could run it on TOR and pay the volunteers via bitcoin.
Donations are weighted depending on the amount donated, people could vote for the target they most wanted addressed first. The balances could just continue to grow until a volunteer accepts the job. Of course it would have to be a COD service and some sort of clear proof would be required, but it's certainly not outside the realm of possibility.
You could probably even get corporate sponsorship.
"[...]and 97% less complexity than traditional servers."
Wait, what? How in the world did they measure this? I'm seriously curious as to this dubious number.
If I mod you up, it doesn't necessarily mean I agree with what you've said, sorry.
Show me all the ARM Cortex CPU's with hardware virtualization support and ECC memory controllers.
I'll absolutely second this - HP's servers kick ass, quite frankly. They've had a few pretty major problems in recent years (P400 and P800 array controllers were absolute pieces of shit from a reliability standpoint, and the P410 STILL doesn't work quite right with SATA drives, though it rocks with SAS disks), but overall the engineering that goes into HP servers puts them well ahead of their competition, from what I've experienced. I've used Dell, IBM, white box, and HP, on the scale of "hundreds to thousands" of each brand, stretching back 10+ years.
The HP's have been more reliable, more configurable, more robust (yes, this is different from reliable), more manageable, and FAR better supported. There's a reason companies pay a premium for HP hardware, and it's because it pays for itself many over during the life of the hardware.
There are companies and applications that don't need that kind of reliability and run on shoddy white-box hardware... think Google, Facebook, etc. There are others, particularly stateful services like telephony and conferencing, that depend on reliable hardware. For those like that, servers like what HP provides will always be in demand. So long as HP maintains their focus on engineering in the server space, they won't be going anywhere soon.
Nazi is a proper noun and thus must be capitalized.
Certainly. But attributive words preceding a proper noun are also capitalized. So it is "Grammar Nazi," not "grammar Nazi."
I can't help thinking here that HP screwed up their courage, took a deep breath, and did the wrong thing. Seems to me the prevailing argument ended up as "PC compatible" thus going with the weak-and-hot atom. But these servers are all going to be running Linux, so where's the argument for PC compabitility?
When all you have is a hammer, every problem starts to look like a thumb.
"Moonshot" refers to their business strategy. This is a 'moonshot', high-risk, high-reward, but more than likely to just go into the crapper like pretty much everything except their calculators and printers.
Sleep your way to a whiter smile...date a dentist!
I doubt they would pay to have a post go live at midnight.
I don't know about Slashdot's priorities when it comes to deciding what makes the cover, but I submitted in good faith.
I'm not a fan of HP by any means & we have truck-loads of their servers at work. The concept for this sounded interesting & maybe there is a place for it in the 'cloud'.
What we really need at work is big kick-ass servers like IBM's new Power 7 machines (IBM - please direct debit my account ASAP)
I think SOC for data centers makes a lot of sense.
Microcontroller and SOC tech is still catching up to current CPUs, but they have a major advantage of cramming just about everything on the mobo into the chip.
In a decade or so we may well be looking at today's data centers the way we currently look at ENIAC.
more robust (yes, this is different from reliable)
What is the difference between robust and reliable?
"First they came for the slanderers and i said nothing."
Ok, mod me into oblivion if I'm wrong, I guess every blue moon there is someone with a 5 digit id as a "first time submitter" ;) I was just rather surprised they posted a story with a link to their own article that was posted a few hours before the referencing one... last couple of times I had a story posted it was at least a day after I submitted it. And of course the HP AD I SAW next to it didn't help the situation one bit...
nazi's.
Have fun.
Calling someone a "hater" only means you can not rationally rebut their argument.
Considering the fact that only a small percentage of even IT people understand just how much server horsepower is required for many typical tasks that environments require of them, I don't expect a huge demand for these Moonshot servers. The specs however are very well suited for many applications used in small to medium sized businesses. And when you get to those who would see appropriate use for these, the price of the chassis is very ugly.
atoms run linux just fine.
Subject: I want one or two
Start with a HP Proliant Microserver 40L: all in all, a 4 bay non-hotswap low power home NAS for about $200 (HDD-es not included) - or make it a media center, or whatever you fancy at that spec.
I reckon you can have one or two at that reasonable price.
Sounds sweet!!
Biggest advantage... the microserver does not sound at all... it's virtually silent
Questions raise, answers kill. Raise questions to stay alive.
if you've done nothing wrong you have nothing to fear.
a third?
maybe we need to construct an Ark?
Is this 330 coupled with an NVidia ION? Because, I have a Atom 330 with Ion chipset and an Atom 525 with the integrated Intel graphics. The 330 is much snappier as a desktop system than the 525, even though it should be less quick. Personally, I think the low-voltage low-end Celerons are a much better deal than anything Atom and I've been an Atom fan for a long time.
Ahhh...the great dumpster continuum. Many a free computer will be found there. -- sowth (748135)
Well the Cortex A15 has both.
Yeah, HP servers weren't bad after they bought Compaq and pretty much abandoned the old line of HP servers.
Learn to love Alaska
I think his argument boils down to "ARM runs linux better".
Whether that is true or not is up for debate.
Slashdot social media options: AIM, ICQ, Yahoo, Jabber and Mobile Text. Why no MySpace?
"[...]and 97% less complexity than traditional servers."
Wait, what? How in the world did they measure this? I'm seriously curious as to this dubious number.
"Now with 67.3% more dubious numbers than traditional advertising copy!"
"Little does he know, but there is no 'I' in 'Idiot'!"
:-) 1st time accepted, 2nd or 3rd submission I think. Sad how the quality of stories seems to have declined over the last few years, but the quality of the posts & the turns they take still sometimes surprise me. Without their seasoned contributors, this site would fade away in no time. I hope the people running it recognise that.
mid-afternoon in GNU Zealand :-)
"Hi, yeah, could I get a number 2 with a coke? Oh, and large fries. And can you reduce the complexity on that? By how much? I don't know, 100%? Oh, you can only do 97? Ok fine, I'll take that. Oh, and a chocolate shake."
Nops, it's a standard Intel motherboard with a I8294G graphic chipset. Barely acceptable, but it does the job. And, as you said, as a desktop machine it's a pain in the mouse's ass.
I had a harsh time, however, until I manage to install and configure the correct drivers and codecs. Win7, as it's installed, does a shitty job on the Atom 330. You need to use the Intel network driver (and turn off all hardware aid!), and do not forget to install the Atom 330 optimized codecs - otherwise you will not be able to see 1080i video.
I also installed a Soundblaster 5.1 PCI soundcard, with the correct drivers.
And that's all.
The machine is running almost 24x7 for 2 years and something, and I have no (many) regrets. My power bill lowered enough to spend some money on yet another 4T of storage for multimedia with the savings not much time ago.
Lisias@Earth.SolarSystem.OrionArm.MilkyWay.Local.Virgo.Universe.org
robust = when it breaks, it can be fixed more easily
reliable = doesn't break as often
it's gonna break either way
They break in predictable ways. All equipment will have failures. Do you want to spend the time swapping power supplies and hard drives that tell you when they have failed, or figuring out what the fuck is the matter with this broken box?
A good example is something I've experienced multiple times. A hard drive in an array fails hard. In a Proliant, you get a red light and the machine keeps running. In a Dell or IBM, it takes down the whole disk bus and you have to take time pulling individual drives and reenabling the array until you figure out which drive is the bad one.
And we all know that datacenter servers are used only for running the front-end of 3D games. Oh no, this product must be DOOMED!
Alternatively, it could be also QUAKEd.
Ezekiel 23:20
The one potential spoiler for SoCs is virtualization.
Sure, the motherboard of your generic dual xeon/opteron box looks a bit untidy(and I suspect that we'll see further integration here, and already have seen some, goodbye discrete northbridge...); but if you divide the number of wasteful little discrete packages across the number of VMs the machine is running, it starts to look a whole lot better.
This isn't a 'bah, integration, it'll never happen!', it has been happening fairly steadily in PCs more or less since IBM defined them and Compaq produced a non-copyright-infringing competitor. Discrete option cards gradually get eaten by motherboards, and once it's an expected motherboard feature, the Northbridge or Southbridge usually engulfs it. More recently, most of the northbridge has been eaten by the CPU. Full SoC-level integration seems unlikely for the moment because PoP RAM severely limits CPU thermal envelope and total system RAM, and because certain specs still vary enough by use case that it isn't economic to go one-size-fits all; but integration proceeds apace elsewhere.
True. In the case of web severs serving static files, the CPU sits around waiting for the hard drive to send a file to the network card. Partly because CPU speeds have increased by 40X while hard drive speeds have only doubled, for many workloads the CPU isn't the limiting factor and an Atom would work just fine.
When I read the article I was harkened back to my days working with an IBM 390 mainframe. Massively parallel blah blah. It seems the pendulum does swing back periodically....
Well, of course they can do more, but how much more? A lot of the OS and application software that runs in my server cabinets is licensed per processor. Seems that the Atom would be a bad fit in that particular scenario.
Referring to Orbiter as a 3D Game (especially at 1280x1024x16) is like going to Maxim's and bringing your own happy meal. Call me when you can play Bioshock Infinite on, at least, medium at no less than 30fps in HD resolutions.
And even their printers are crap now days.
There must be balance in the force, young padawan.
If either side gains, then the retaliation swings the line of power back and forth like a pendulum.
And before you know it there's a clone war, brothers kissing sisters and lil furry creatures saving the galaxy with rocks.
We must hold the line firm!
not really, a mainframe of S390 era had very few cores. mainframe architecture isn't about a bunch of systems network connected all running separate OS instances. It's instead a "star architecture" of processors connected to "peripheral processing units" to offload IO work.
subject says it
Shooting the moon
Google it. Or Bing it for the free Redbox rentals or Amazon gift cards.
Heavens for Betsy, did you start a sentence with the word "and"? What is that sound? English teachers attempting suicide with a large ice picks in their eyes.
When you're talking about 6watt CPU's packed together as close as possible, I think PoP RAM makes perfect sense.
Just think how much space is wasted by RAM sockets PCB traces and how much power is wasted driving high frequency data lines between chips.
Is it 64 bit?
I ran Linux on mine. I guess Windows would be better, as the Intel drivers for Atom aren't open source. With the ION it was just fine, due to the (closed source) NVidea drivers. They served me well, but due to space issues, I had to fall back on having all my gear in a backpack. The two Atom machines are now somewhere in a closet at my parents. I'll surely find a use for them someday...
Ahhh...the great dumpster continuum. Many a free computer will be found there. -- sowth (748135)
megabuy.com.au
Questions raise, answers kill. Raise questions to stay alive.
Is it 64 bit?
Internally yes, with some instructions processing 128-bits. The address buss is 40-bits wide, limiting? physical memory to 1TB
Certainly, for very low power processors, PoP makes sense(at least until you hit the ceiling on how much RAM you can actually fit on top of a very low power processor, the 8GB SODIMMs in these little HP boxes generally have 16 little BGAs, all not significantly smaller than the CPU itself, on the card). My question is whether, given how easy it is to slice a larger CPU into smaller virtual CPUs, the 'lots and lots of teeny CPUs' architectural strategy is actually a good one, or just a good one until it finally scares Intel into not overpricing Xeons, especially the ULV ones.
That's like saying the Atom is 128bit, because it has 128bit SSE registers.
The Cortex A15 is an ARMv7. All registers are 32bit, apart from the SIMD NEON ones which are 128bit. Sort of like the 128bit registers in SSE.
It's a 32bit core with a 32bit memory bus and a single 32bit 800mhz memory interface.
ARMv8 is 64bit. It has 64bit registers
That's the Cortex A53 and A57.
Why didn't you just say "The Cortex A57" instead of spinning lies about the A15?
It's a lot easier to slice up a large CPU in to several smaller virtual CPU's than it is to use several smaller CPU's to emulate a large CPU.
In my experience they completely kick Dell's butt, and give IBM a real run for their money, at much lower cost. I evaluated a ProLiant Gen8 and the manageability features were pretty impressive. The thing can update it's [sic] firmware and send SNMP traps, etc, from bare metal, without an OS.
Those have been fairly standard service processor features, actually, for a number of years. The SNMP trap thing is slightly annoying, in that everyone seems to assume that some application that can handle traps is available. The latest iLO 4 firmware introduces the ability to syslog (which is somewhat broken still - and something Sun did years ago) and send email (also done by Sun years ago). Unlike IBM and Dell and Cisco's UCS servers, HP Gen 8 systems with iLO 4 >1.05 have serial consoles that work right out of the box for bootstrapping (also done by Sun years ago). iLO also has the ability to mount an ISO image over the net via an HTTP URL, which is quite handy at times. The HP Gen 8 systems (with iLO 4) really are a distinct step above the G7 systems in a number of ways, eg. PXE booting doesn't inexplicably force the console to 115200 bps, lack of the bizarre MCE issues that some G7 systems had, much better iLO, etc. Mid 2012 they finally came out with 25SFF / 10SFF chassis so that a decent number of internal disks can be provisioned, with SAS expanders embedded in the backplane. And most notably, at one point I complained to HP about their Smart Array HBA's not being able to do 3-way mirroring. I dunno for sure if they took my RFE personally, but the Gen 8 series of HBA's actually has that ability, something valuable I had not seen on products from Adaptec and LSI.
Granted, HP had some crappy CEOs, and on the low-end consumer stuff they race to the bottom with everyone else, but their servers are serious and arguably industry-leading.
Given the current product line, sure -- but the G5 / G6 / G7 systems had their share of suckage, including iLO systems that were painfully underpowered and sluggish. I demo'd a G5 system at one point and was incredulous that the PCI riser was attached to the chassis lid, such that opening the lid entailed ripping cards out of their slots. I think G5 and earlier systems also didn't come with the serial console working out of the box, which was a non-starter. HP, Dell, and IBM systems still share a bit of anachronistic suckage: the serial console is a DB9, straight out of 1990. Sun's systems have used RJ45 connectors for *years*. HP refuses to even include or even spec an adapter for their DB9 consoles -- we had to try a bunch of different models to see what worked, then I bought 50 of the things.