Blades and Virtualization Aren’t Mutually Exclusive: Part Three, IBM Traditional Expansion Options

By Aaron Delp

UPDATE:  I have made some corrections to the 10GB and iSCSI sections based on feedback.  Thank you!

Welcome to part 3 of my series on blades and virtualization. I will be breaking the expansion articles into four parts: traditional IBM options, traditional HP options, virtualized IBM options, and virtualized HP options. By “virtualized options” I mean the IBM and HP blade I/O virtualization solutions (like IBM Blade Open Fabric Manager and HP VirtualConnect), not a server configuration that will run virtualization. Today, we will be covering IBM traditional options.

Let’s take a second to revisit a portion of our hardware standard we have been using for the power series and make a few additional notes. Here is the spec again:

  • A minimum of 32 GB memory – As many have commented and my own studies have shown; memory is usually the first limiting factor. The more memory you can install in the system, the better.
  • 2 on-board NICs + some combination of expansion cards (FC, iSCSI, NIC, 10Gb)
  • 2 x 146GB, 10k SAS drives (except the IBM HS21XM blades which only support one HD)

I will get into the expansion options in a second, but let me start with the memory since this is the area most people are concerned about. With the recent addition of 8GB DIMMS to the HS21XM product line, the HS21XM now supports 64GB of memory. The IBM 3550 supports a max of 32GB and even the IBM 3650 only supports 48GB. If memory is your concern, the IBM blade is actually the BEST fit for virtualization.

Another area of concern is the fact that the HS21XM only supports one hard disk. If you would like to boot the OS from a local disk, this can be an issue. IBM also offers the dual solid state drive in a RAID-1 configuration but both platters are mounted in the same enclosure so if you lose one, you need to replace them both. We know because it happened to one of our customers. In the end that doesn’t make sense either. I don’t see the actual ESX build as critical in a VirtualCenter environment so I usually don’t worry about it excessively.

The expansion cards can get a little tricky depending on the configuration. For virtualization I see four possible options to fill out the blades to meet our needs. They are FC, 10Gb, iSCSI, and NFS. We will fill any remaining expansion with Ethernet 1GB ports to provide as much connectivity as possible.

Above is a picture of an IBM BladeCenter H Chassis from the rear, highlighting the expansion bays. Bays 1 and 2 connect to the onboard Ethernet so the switches must be Ethernet compatible. Bays 3 and 4 connect to the first expansion card on the IBM Blade, often called the CFFv form factor. The v in CFFv stands for “vertical” meaning this card will talk to the vertical expansion switches.

Bays 7-10 often cause much confusion.  They are positioned horizontally in the chassis and connect to any CFFh (h for horizontal, get it?) card on a blade. A number of different switches can be inserted into these bays. The first one is the easiest one, the 10Gb switch.  The first switch would go in bay 7.  The second would go in to bay 9.  Each switch would connect to a port on the dual port 10Gb card.

UPDATE: You can also do a 4×10GB configuration by adding 2 more 10GB switches in Bays 8 and 10.  You will then need the 4 port 10GB card instead of the 2 oprt 10GB card.  I have updated the charts below to reflect the MAXIMUM number of connections.  You can of course go with less.

The second option for Bays 7-10 involves something called the MSIM (Multi-Switch Interconnect Module). This option will allow the FC, iSCSI, and NFS configurations. The MSIM is an adapter that divides the High Speed Bays into Bays 7 to 10 shown above. To take advantage of all four bays, two MSIM’s divide both High Speed Bays and you will need a four port CFFh card to map to all four bays.

The above information yields the following switch and blade expansion card combinations:

Blade with 10Gb & 4 NICs (no MSIM’s – 4 NICs and 4 10Gb per blade):

Dual On board NICs

Bays 1 & 2 (Ethernet switches)

Dual Port Cffv NIC

Bays 3 & 4 (Ethernet Switches)

Dual Port 10Gb Card

Bays 7,8,9,10 (10Gb switches)

 

Blade with 10Gb, FC & Eth (no MSIM’s – 2 NICs, 2 FC and 4 10Gb per blade):

Dual On board NICs

Bays 1 & 2 (Ethernet switches)

Dual Port Cffv FC Card

Bays 3 & 4 (FC Switches)

Dual Port 10Gb Card

Bays 7,8,9,10 (10Gb switches)

 

Blade with FC Option #1 (requires 2 MSIM’s – 2 FC and 6 NICs per blade):

Dual On board NICs

Bays 1 & 2 (Ethernet switches)

Dual Port Cffv FC Card

Bays 3 & 4 (FC Switches)

Quad Port Cffh NIC

Bays 7,8,9,10 (Ethernet switches)

 

Blade with FC Option #2 (requires 2 MSIM’s – 2 FC and 6 NICs per blade):

Dual On board NICs

Bays 1 & 2 (Ethernet switches)

Dual Port Cffv NIC

Bays 3 & 4 (Ethernet Switches)

Dual NIC and Dual FC Cffh Card

Bays 7,9 (Ethernet) & 8,10 (FC)

 

Blade with NFS or software iSCSI (2 MSIM’s – 8 NICs per blade):

Dual On board NICs

Bays 1 & 2 (Ethernet switches)

Dual Port Cffv NIC

Bays 3 & 4 (Ethernet Switches)

Quad Port Cffh NIC

Bays 7,8,9,10 (Ethernet switches)

 

You will notice that only 3 of the 4 options are given. Hardware-based iSCSI is missing. That is because IBM has a hole in their portfolio. The IBM iSCSI card is the older form factor that if used on an HS21XM blade it doesn’t allow any other card. So, you end up with 2 NICs and 2 iSCSI. This isn’t acceptable and isn’t a valid option.

So what do you think? I believe that any of the above configurations will suit virtualization nicely given a proper design. I look forward to your thoughts as we continue on to the next article, HP traditional expansion.

Tags: , , , ,

  1. Bill B’s avatar

    Good analysis.
    I checked the list price on an HS21XM blade with 2 E5450 processors, 64GB memory, 6 X GbE and 2 X 8GB FC connections - $15,784
    Same blade with 32GB memory - $9,520

    Do you think the extra 32GB of memory is worth it in terms of being able to support that many more VM’s on the blade?

  2. Justin’s avatar

    Love this series!!

    Two things, you can go with mirror Solid State drives in the HS21XM. However, I haven’t been able to justify the increased price in retrospect to what i gain with HA and DRS functionality.

    Why the focus on the 3550 and 3650? Are you trying to stick with the 2U/1U form factor?

  3. adelp’s avatar

    Bill B - For that question you get my standard engineering answer - “It depends”. :) By that I mean hopefully a virtualization TCO study was done and you will know that answer for yourself. I have done a number of VMWare server consolidation studies and we often see advantages up to 32 GB but at some point the memory vs virtual machine density levels off. Only a study of your unique environment can tell you where the sweet spot is. It might be 32GB, 48GB, 64GB. I even had one customer that the peak was 16GB (they were CPU bound, not memory bound).

    Also, have you considered Kingston memory in the servers? The 8GB Kingston DIMMs work great in the HS21XM and are supported. They might be cheaper for you and you will stay under IBM support. We have some customers doing this without any problems (they actually ran them a long time ago before they were supported but that is another story).

  4. Wayne I’s avatar

    Aaron, just a minor point, but we run or HS21XMs with the dual solid state drive in a RAID-1 config. We have experienced a “platter” failure and were able to replace just the single platter. It appears that both the platters and the daughter card they plug into have seperate FRUs. However I do agree with your point that the ESX build is not critical as we deploy our ESX blades using one of the deployment appliances. 15 minutes and you have a new ESX server!

    Wayne.

  5. adelp’s avatar

    @Wayne I - That s good to know! We had a customer fail one and IBM Support told them you HAD to replace the entire assembly. I was unable to find a FRU for the subpieces of the assembly. The customer ended up replacing the whole thing and we reloaded ESX. Do you happen to have the FRU’s? I would love to file them away and post them on my blog. Thank you!

  6. HP BladeSystem Team’s avatar

    Scott, great write up on I/O expansion for virtualization.  We agree – historically, barriers to virtualization in a bladed environment have been limited connectivity,  high networking costs, fixed network bandwidth and insufficient memory and excessive power use.  We look forward to the rest of this series.

  7. slowe’s avatar

    Aaron gets all the credit for this series…and there is more yet to come!

  8. Net Pro’s avatar

    On our HS21XM blades the FRU for a 15.8 GB SATA drive (SSD) with a transfer speed of 1.5Gb/s is: 43W7609.
    The FRU for the “adapter” where the two SSD drives plug into is: 39M4583.
    These days IBM offers an SSD drive with double the capacity but rumor has it that it cannot be configured as RAID-1, not sure if it is true.

  9. william bishop’s avatar

    Aaron, I’m told by an IBM rep that the 8gb kingston memory modules aren’t supported by IBM….Have you got a support sheet on them?

  10. adelp’s avatar

    Net Pro - Thank you very much for the part numbers!!!

    William Bishop - I spoke to our Kingston rep and he send me the written statement of the agreement between IBM and Kingston. It covers all products. Is your e-mail on the comments valid? I’d be happy to send you the soft copy.

  11. william bishop’s avatar

    Yep, valid.

  12. BillH’s avatar

    One correction - The 10Gb Ethernet switch only takes one switch slot. The first one would normally go in Switch Bay 7. A second 10Gb Ethernet switch module would go into Switch Bay 9.
    http://www-03.ibm.com/systems/bladecenter/hardware/openfabric/ethernet.html

    Addition - If you want more connectivity, there is a four port 10Gb CFFh card that can be used with up to four 10Gb Ethernet switches in the chassis. (this is in addition to the base two 1Gb Ethernet ports, and two ports (FC or Enet) added with a CFFv card.)

    Addition - There is recent higher bandwidth ‘combo’ CFFh adapter (dual Gb Ethernet + dual 8Gb FC)

    Comment - iSCSI is supported very well with the basic Ethernet NICs, so no iSCSI specific HW adapter option is being offered.

  13. adelp’s avatar

    @william bishop - I sent you an e-mail with the statment. Let me know if you need anything else.

    @BillH - Good points all around. I will correct the article on Bays7-10 for 10G. I was not aware of the 4×10G solution. I have questions out to some people in IBM on it because I don’t see how you can push that kind of bandwidth through the chipset. You can’t do it on a 3850 and still be supported without a SPORE. I am amazaed that is a supported config but I have been wrong before!

    Yes, 8GB is now out so anything FC above can really be 4GB/8GB.

    My intention with iSCSI is VMWare. I should be more clear in the article. I will correct that. There is no hardware iSCSI support for the other NICs. The problem with the current IBM solution is HP now has a card out that fits with everything else and IBM doesn’t.

    Thanks again for all the comments!

  14. BillH’s avatar

    I couldn’t mention the new blades prior to announce (of course). You are correct that older servers could not deliver full bandwidth dual-port 10Gb Ethernet over a PCIe x8 bus segment. (Realistic bandwidth would top out around 13Gbps in each direction, shared across the two ports.)

    The full bandwidth of four 10Gb Ethernet ports on each blade does require the newest/fastest chipset. The HS22 blades have the Nahalem Intel processors, which now support PCI-Express (PCIe) Gen2 I/O. This provides double the I/O bus speed when a Gen2 device is connected. (I haven’t seen published numbers yet, but I expect to see well over 25Gb in each direction from a Gen2 x8 PCIe bus segment.)

    The 4-port CFFh I/O card uses two dual-port Broadcom 10Gb Ethernet chips. It is a small hardware detail, but a significant performance benefit, that each of the two chips has its own 8x PCIe bus segment. This is unique to the IBM BladeCenter CFFh architecture. Unlike standard PCIe adapters that provide just a single PCIe bus segment. A standard PCIe adapter would require the addition of a PCIe switch to split the available bandwidth from the server and allow connection of the two Ethernet chips. The CFFh card can provide a large single PCIe bus (a x16 Gen2 bus segment capable of a datarate somewhere over 50Gbps!) or multiple full PCIe bus segments in various combinations of x8 or x4 PCIe widths, all capable of Gen2 PCIe speed operation - with newer , Gen2 capable adapter cards. (Of course, PCIe Gen2 is fully backwards compatible with existing PCIe Gen1 devices. The links autonegotiate to the right speed, without any manual intervention required.)

  15. Aaron Delp’s avatar

    BillH - Thank you as always for the informative comment!