Blades Made Simple - Latest Comments

Re: New Use for Blade Servers? HP E5000 Messaging System

camberleyb — Fri, 13 Jan 2012 11:22:05 -0000

I know a bit late on commenting, but was just pointed to this. We had the E5000 in our lab for the purpose of looking at ease of use and doing some comparisons. Our experience was very good. Regards, Camberley Bates, www.evaluatorgroup.com

Re: HP Flex 10 vs VMware vSphere Network I/O Control for VDI

Andrew VanSpronsen — Fri, 06 Jan 2012 15:00:31 -0000

Agree 100%, we shifted away from Flex-10 network bandwidth controls to NetIOC about a year ago. FlexFabric is also a non-starter as we are NFS.

Re: HP Flex 10 vs VMware vSphere Network I/O Control for VDI

Ken Henault — Thu, 05 Jan 2012 15:40:30 -0000

I have many customers that have deployed FlexFabric with NetIOC. This gives you the best of both worlds.

Advantages:

·
Cost savings and
flexibility of FlexFabric

o   This is doubly important when you have enclosures that
mix virtualized and non-virtualized servers


Simplified cabling

·
Dynamic bandwidth
management from NetIOC

·
Define ESX
networking at the cluster level instead of the host level

and more that I can't think of right now.

Disclaimer: I work for HP, but all opinions are my own

Re: HP Flex 10 vs VMware vSphere Network I/O Control for VDI

Eric van der Meer — Fri, 16 Dec 2011 11:25:39 -0000

Thanks for sharing Dwayne, good post!

But Julian is right:
"The HP 6120G/XG Blade Switch provides sixteen 1Gbdownlinks and four 1Gb copper uplinks, two 1Gb SFP uplinks along with three 10Gb uplinks and a single 10Gb cross-connect"

Instead you can choose the HP 6120XG Blade Switch switch, this one has:
"HP 6120XG Blade Switch provides sixteen 10Gb downlinks and eight 10G SFP+ uplinks" but I don't know the price of that module... and btw, what does list prices say ;)

Re: HP Flex 10 vs VMware vSphere Network I/O Control for VDI

dlessner — Fri, 16 Dec 2011 11:09:43 -0000

Julian, seems your are right about the down links, I guess that's one way for forcing you to buy the Flex-10. No free meal ticket with 10Gb I guess. Still like the config for VDI even with Flex-10.

Thanks for your comment

Re: HP Flex 10 vs VMware vSphere Network I/O Control for VDI

Julian Wood — Fri, 16 Dec 2011 10:45:35 -0000

Great post, Dwayne, yes, NIOC can be used very successfully to more dynamically carve up the Nics which is certainly more flexible that Flex-10.

I think the HP 6120G-XG only presents a single 1Gb downlink to the blade so using this you are limited to 2 x 1Gb with two switches. Flex-10 provides you with 2 x 10GbE

Re: Nutanix Cluster: Disruptive to Blade Server Market?

agismaniax — Wed, 30 Nov 2011 04:11:12 -0000

after browsing into nutanix website, it looks like they were using supermicro's 2u twin2 product - 2U chassis with 4 server nodes inside it. each node is equipped with 2x intel xeon, enough ram, 1x fusion-io ioDrive, 1x 2.5" SATA SSD and 5x 2.5" SATA HDD. the main different is the software inside each of nutanix box. nice packaging!

Re: 4 Socket Blade Servers Density: Vendor Comparison (2011)

agismaniax — Tue, 29 Nov 2011 07:30:07 -0000

how about fujitsu primergy bx900 s1 and supermicro superblade? in single chassis, fujitsu can hold 9 quad-socket intel based blade servers, whereas supermicro can hold 10 quad-scoket amd based blade servers.

Re: Gartner Releases Magic Quadrant for Blade Servers

Diego Tomazzoni — Sun, 27 Nov 2011 12:07:24 -0000

Dell and HP dont have redundant back plane!

Re: Why Blade Servers Will be the Core of Future Data Centers

Michael — Sun, 20 Nov 2011 20:12:59 -0000

From my perspective we have to look more from a mathematical and physical and chemical aspect.
At the momen you have 4 core elements. CPU, Memory, Network and Disk.
Disk is at the moment the slowest component.
This has changed with satadimm.
http://www.vikingmodular.com/p...
The next big thing comes with Reram, FERAM and Memresistor in 2013. This are 100 to 1000 faster than ssd.
http://www.nvnews.net/vbulleti...
HP, Hynix, Samsung and others are now developing Flash NAND successor called
ReRAM that are 10 times faster than 22nm NAND, able to double the memory space
cheaper than NAND, use half power consumption than NAND and can write more than
1 trillion write cycles per sector. SSDs with ReRAM will launch in 2013, it will
replace both Flash NAND SSDs and HDDs as ReRAM now far more reliable than either
MLC and SLC Flash NAND and it finally can be use for very heavy users.
HP's memristor reality in 2013, can revolutionize SSD and RAM market
http://www.nordichardware.com/...

http://forums.vr-zone.com/news...

The ReRAM or resistive random access memory chip consumes less power and is
capable of writing data 10,000 times faster than NAND flash memory

So this the means if the disk is elminated , we will have 3 elements left. So it is 3! (factorial).
RAM and ReRAM will may be use the same Dimmslot.

As you can see with the viking satadimm you can build extrem small micro servers.
So than the dimm is the biggest component in the computer.

The network will than be the slowest component.
http://www.usenix.org/event/ho... (Page 6 Nic is moving into the cpu).
http://www.scs.stanford.edu/~r...

Re: Why Blade Servers Will be the Core of Future Data Centers

j_e_w — Sat, 05 Nov 2011 08:59:13 -0000

Dear Kevin,
first thank you for that clear and simple illustration and language you are using through out your blog. It is a pleasure.

I totally agree with you in terms that the future should look like this. I do not understand, why DELL, IBM and HP not drive that kind of development faster or - let's say it better - why they don't relaese it (sure they all have already working prototypes of stuff like this). Not only the user would get greater integration and simplification of his datacenter. But the vendor would tie the customer even tighter to his brand, since all components are somehow backplaned which virtually locks the buyer within the particular brand.
I also dear of greater SAN integration within baldechassis offering wider storage capabilities as IBM BC-S..;)

Thanks and best wishes from Munich
Jew

Re: Why Blade Servers Will be the Core of Future Data Centers

j_e_w — Sat, 05 Nov 2011 08:59:03 -0000

Re: Dell Network Daughter Card (NDC) and Network Partitioning (NPAR) Explained

Brad Hedlund — Fri, 04 Nov 2011 11:08:23 -0000

The "big deal" with the Dell approach is that NIC partitioning is *independent* of the upstream switch. Both the HP and Cisco approach lock you in to their switches (Flex-10, or Nexus).

Cheers,
Brad

Re: Dell Network Daughter Card (NDC) and Network Partitioning (NPAR) Explained

tonybourke — Sat, 29 Oct 2011 20:24:34 -0000

Great write-up and great for getting familiar with Dell's offering. However, like HP Cisco has also been able to carve up NICs for a few years with UCS. The M81KR and VIC1280 cards can crave about 50 (M81KR) or 100 (VIC1280) abstracted NICs/FC HBAs and assign bandwidth/QoS to them.

Here's a great writeup by Brad Hedlund (who's just joined Dell Force10).

http://bradhedlund.com/2010/09...

Still, cool stuff all around.

Re: Why Blade Servers Will be the Core of Future Data Centers

Adam Richardson — Fri, 28 Oct 2011 08:59:55 -0000

Actually its more like the SuperDome 2 architecture rather than 3par, you need a crossbar rather than a meshed architecture to make scalability work to the level this is aiming at. You need to be able to "partition"resources off at a physical layer.

That second diagram is the original SuperDome circa 2000. Only difference is that its still assuming CPU and memory live on the same blade, take it one step up and assume memory, CPU, storage (actually with things like memristor or PCM the memory AND storage could be the same thing) and I/O are seperate bladed modules all linked by a common interconnect and then throw in the ability to start zoning/partitioning the communication between them and you have the true converged architecture. Plus you've just eliminated the need for a hypervisor.

Now have a look around at which vendors could actually execute on that ..............

Re: Why Blade Servers Will be the Core of Future Data Centers

Chris Fricke — Thu, 27 Oct 2011 20:02:01 -0000

Ya know... HP is not too far off from this actually. If you took the backplane architecture of the big 3Par, added some physical compatibility to plug in compute, and whisk in some secret sauce to manage it all... maybe?

The question is: would it sell?

I wouldn't buy it. It's too big.

Re: Why Blade Servers Will be the Core of Future Data Centers

Daniel Bowers — Tue, 25 Oct 2011 16:56:08 -0000

Rackplane looks more like ATCA to me than rack-mounts. Or maybe a mainframe.

Clever idea, though I have two objections:
1) It's *too* flexible. The rack midplane would have to be high-speed, low-latency, and resiliant. It would need to allow any device to go in any bay. Those kinds of cross-connects exist (e.g. in a Superdome, Fujitsu M-series, etc), but they'd be expensive and over-kill for most users.

2) I don't like the idea of compute nodes being just "processor plus memory". If you start adding flash storage as an 'accelerator' accompanying memory (like what was mentioned for Dell PowerEdge 12th gen at DellWorld), then the "memory" part might be something drastically different (and bigger) than a handful of DIMMs. Alternatively, maybe future servers will physically seperate compute cores from main memory, so maybe the two don't always belong together.

Re: Why Blade Servers Will be the Core of Future Data Centers

Joe Onisick — Tue, 25 Oct 2011 14:18:28 -0000

Kevin,

Great blog I enjoyed the thought experiment. After rereading it I believe you may have just invented rack-mount servers!

Joe

Re: Why Blade Servers Will be the Core of Future Data Centers

Joe Onisick — Tue, 25 Oct 2011 14:17:41 -0000

Kevin,

Great blog I love the thought experiement. After reading through this twice I think you may have just invented rack-mount servers.

Joe

Re: Dell Network Daughter Card (NDC) and Network Partitioning (NPAR) Explained

Kevin Houston — Wed, 12 Oct 2011 07:57:46 -0000

While carving up network ports into virtual NICs has been around for a while, the #Dell NPAR capability is on the NIC chip and does not rely on any specific module. The ease of use, and the ability to use with any of the 10Gb Ethernet Modules offered by Dell makes it more appealing (to me at least) then other competitive offerings. Thanks for reading, and thanks for your comments!

Re: Dell Network Daughter Card (NDC) and Network Partitioning (NPAR) Explained

Geert — Wed, 12 Oct 2011 05:31:49 -0000

What is the big deal about this ? HP has this already for years: Flex-10

Re: Dell FlexMem Bridge Helps Save 50% on Virtualization Licensing

Scot — Mon, 03 Oct 2011 11:41:48 -0000

Hi Everyone, with the latest vSphere 5.0 release. Does the new license
model hurt Dell's FlexMem virtualization licensing costs now that it
charges for increased memory capacity for vms? Does this diminishes the
FlexMem value-add?

Re: Why Are Dell’s Blade Servers “Different”?

Andreas Erson — Sun, 02 Oct 2011 17:56:06 -0000

Actually it wouldn't be possible to have a NDC with 40GbE-ports since the LOM is "only" connected by 1-2 lanes to each of the two fabric A I/O-modules. The mezzanines are connected by 1-4 lanes to each of the two fabric B (or C) I/O-modules.

Utilizing 10GBASE-KR the M1000e is designed to handle the following max configuration for a half-heigt single-width blade:
LOM/NDC = 4 port 10GbE (using two 10GBASE-KR lanes to each fabric A I/O-module)
Mezzanine, fabric B = 2 port 40GbE (using four 10GBASE-KR lanes to each fabric B I/O-module)
Mezzanine, fabric C = 2 port 40GbE (using four 10GBASE-KR lanes to each fabric C I/O-module)

My guesstimate is that the next step in the M1000e will be 4port 10GbE mezzanines and a corresponding new I/O-module perhaps called M8048-k (32 ports internal, 16 ports external). If you take a current 4 port 1GbE Intel the two dual port gigabit ethernet controllers are 25x25mm in area. Intel also produces dual port 10GbE ethernet controllers with the same 25x25mm size. Each mezzanine is connected with PCIe x8 2.0/2.1 to the motherboard giving it a total of 32Gbps total bandwidth (4port 10GbE would be oversubscribe the connection to the motherboard). With the upcoming 12G servers using PCIe x8 3.0 this will increase to 64Gbps which would avoid any "problem" with oversubscription (let me see the server that actually uses this amount of bandwidth). So in essence the only real issue from my layman point of view would be thermal issues.

To further speculate about 2port 40Gbps without oversubscribing the connection to the motherboard we would need to use PCIe which will probably provide 128Gbps if using a PCIe x8 4.0. But now we are talking a couple of years from now.

Anyhow, give me a 4port 10GbE mezzanine with NPAR and a corresponding I/O-module and I would be very happy. :)

Re: About Us

Michael Thomas — Wed, 28 Sep 2011 08:12:10 -0000

There have been a lot of changes in the technology world recently, but no updates from bladesmadesimple.com. Is there an underlying reason for this site kind of dropping off the grid?

Re: Blade Chassis I/O Diagrams

Chris Barrett — Fri, 23 Sep 2011 19:20:33 -0000

You will now need to update your diagrams for the new I/O modules and Fabric Interconnects.

The Chassis now supports up to 160Gb/s (80Gb per FEX) in a port channel configuration. The new 6248 is based on the hardware specifications of the 5548 with 4k VLAN support, 2 us latency and Universal Ports.

In addition the new 2208 I/O module (aka. FEX) module supports 32 internal facing 10Gb ports each!