Slashdot Mirror
HP Claims Their Moonshot System is a 'New Style of IT' (Video)
Didn't we already have something kind of like this called a Blade server? But this is better! An HP Web page devoted to Moonshot says, 'Compared to traditional servers, up to: 89% less energy; 80% less space; 77% less cost; and 97% less complex.' If this is all true, the world of servers is now undergoing a radical change. || A quote from another Moonshot page: "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.' These are software-defined servers. HP claims they are the first ones ever, a claim that may depend on how you define "software-defined." And what software defines them? In this case, at Texas Linux Fest, it seems to be Ubuntu Linux. (Alternate Video Link)
I can see this being useful for approximately ten percent of the market.
There's more! Buy now and receive a second HP MS System for free! Just pay shipping and handling.
Not available in any store.
But Moonshot servers are a couple years old, with a few success stories from HP itself (www.hp.com is fully moonshot-powered) and others. Yes they are efficient, small and easy to run, but they are also quite less powerful than a "traditional" server. Now all they do is release new "cartridges" for the platform. Are we soon to hear about generation 2.0? Maybe at HP Discover?
4.3U? They couldn't have made a reasonable tradeoff to go to an even unit size?
"If you do algorithms, things of that nature, you can run on these systems."
Sold!
4.3U? They couldn't have made a reasonable tradeoff to go to an even unit size?
Maybe they have a 0.7U add-on planned for it :-)
Being in IT sales, I am often required to surf HP's website. Their site is consistently painfully slow. You would think that a company like HP would make sure their servers could serve up webpages faster than a snail.
If I had the money, I'd totally buy it and avoid the cluster****ery that is cloud services.
BUT...
Notice what the average cpu is. Intel Atom class hardware. Or in otherwords, this is designed for doing dreamhost-style weak cloud VPS, so while you may have 45 servers in the box, the net performance is ... well...
The Atom processor picked, S1260 (2 core, 4 thread @ $64.00 )has a passmark of 916
The highest rated is Intel Xeon E5-2697 v2 @ 2.70GHz, passmark 17361
So 19 of those Atoms (38 cores, 76 threads) = 1 E5-2697v2 (12 core, 24 thread @ $2614.00)
One dual E5-2697v2 server is almost equal, and you have 24 usable cores that could be turned into weak VPS servers. Get the point I'm making?
Moonshot might be a better choice for provisioning weak dedicated hosts instead of VPS's (which are inherently weak, even when running on solid hardware, they are still subject to being oversold.) The S1260 is 64$, the E5-2697v2 is $2614, or roughly the cost of 40 of the Atom's. So on paper someone might go "oh look I can can afford an entire moonshot server for the price of a single cpu E5-2697v2 and get twice as many cores, when the single thread performance of the 2697 is a passmark of 1,662 (yes . 181% of the 4 threads of the Atom.)
The thing is, this kind of configuration is better suited for certain tasks, such as a web server cluster front end (where it's rarely the CPU, but the network infrastructure that's the bottleneck) where you can turn on or off identical servers as required, and none of them actually need hard drives connected, they can just be PXE booted from a NAS.
Though I'm worried when I see "software defined" anywhere in marketing, as most virtualization software fails hard when under load (75%CPU.) So maybe a data center that is space/power constrained can see a use for this, but if you're running a high usage website, you're better off optimizing the software stack (like switching to nginx or using Varnish in front of apache httpd +php-fpm instead of leaving it at the highly inefficient httpd prefork+mod_php) than deploying more inefficient servers.
This sounds like a great idea, right? 45 servers in a single chassis? With an OA (onboard assistant) to allow administration of each individual blade. So about 12 months after you've implemented this monster in your production (no downtime) environment a blade fails. No problem you replace the blade. But the NEW blade comes with firmware that requires an update to the OA (entire chassis) and the new OA firmware won't work with the other 44 blades until you update them also. Hmmmm... hey boss, can I get downtime for servers 1 thru 45 to update our blade chassis? No? Ok well I guess we are hosed unless you get TWO chassis fully loaded and cluster between them.
The whole promotion seems to resemble everything from IBM PureServers that were introduced about 2 years ago, but of course lacking any type of performance. At least the IBM servers allowed scaling, higher performance CPUs, integrated disks, etc..
When management and marketing design computers, this is what we get. HP has not really been a technical player for a long time, at least in terms of innovation. Superdome was okay, but Sun E class machines made them look like an old mainframe in terms of usability. Itanium flopped and they never put much into the PA RISC chips after that. Omniback and NNM were great, but required manpower and HP has despised T&M billing for as long as I've worked with them which goes back to HP-UX 9 and VUE days. (I contracted for them in Michigan, because they would not hire direct technical people).
-The wise argue that there are few absolutes, the fool argues that there are no probabilities.
Imagine if they could back-port this work to their current range of x86 blade servers !
They forgot the golden rule of IT. If your company has the #1 worst rated consumer customer support and the #1 least reliable laptops (emachines beat them at desktops) then don't create a brand new technology that people will be hesitant to use. You pretty much have to be the exact opposite. Only the best company can come out with something new, claim "just trust us, it works perfectly and you should use it" and have people believe them. I really hope this finally bankrupts them so I can stop having to put out HP-induced fires at my business. I'ms serious, two dc5700's lit on fire.
Cisco got fr1st post!
The "Civilized World" jumped the shark ca. 1973.
So, like VDI? Because that works SO WELL.
For large sets, this will be our guide even unto death, for the LORD will work for each type of data it is applied to...
We bought some Transmeta-based blades at $LARGE_US_BANK a while back, and they sucked. Hard. Like, don't bother running stuff on them hard. They went back to HP, or in the trash, I forget, and we got real hardware. It looks like HP is reviving the concept of shitty servers for people who don't do a lot with them. Instead of 1 beefy 4U machine, you have a 45-node Beowulf cluster of suck, and most problems ARENT trivially scalable. Or, if your app really is that scalable (or you've spent the time to make it so) then you're a big boy company and you need real iron.
Fail.
I want to delete my account but Slashdot doesn't allow it.
Look at a SuperMicro catalogue from around 2008 onwards or Verari from even earlier.
The company I'm at is looking at some new serious number-crunching servers.. We had a HP rep come in and propose a Moonshot system. The head of IT and I looked at each other and laughed out loud... Moonshot uses ATOM processors. I don't care how many of them you have, we're not using a rack of low-ball processors in our system. Moonshot is a complete joke.
I think they use ATOM processors because it was the only way to get such high density and still be able to get the heat out of the system. It also may be that Intel has a boatload of atom processors to unload with the crash of the netbook market. We told them come back when HP can get a Xeon in each module..
Moonshot is a great concept that was implemented very poorly. I asked the rep what market Moonshot servers are targeting. He wasn't able to answer the question.
We got demoed this 6 months or so ago.
I still fail to see what this buys you over a bunch of regular blades or rackmounts running your virtualisation platform of choice.
It's 10U with 64 almost real servers (haswell xeons) and has integrated storage and networking. You only need to hook up some power and a beffy uplink to it and you're done. And did I mention a rest api to controll it? Works even with openstack baremetal if you want that. Last I heard (two weeks ago) Moonshot is still only at cli. ... too little, too late.
Apollos on the other side, those are worth considering. But Moonshot
it must be a nightmare if the chassis fails