At least, that’s what I thought.  Then I started having people react to our solution and tell me how ‘nice’ it is we actually did something as simple as NIC Teaming.  So I wanted to list out the functionality we offer on our 10Gb NIC part of the solution.

If you look at the traditional NIC function that has been in any server platform now for almost ten years there are a list of basic features you expect to have to ‘just work’  they can include:

• NIC Teaming
• VLAN Support
• Wake on LAN Support
• Serial over LAN support (IBM BladeCenter Feature)
• Software iSCSI Initiator and iSCSI Software Boot Support
• Preboot Execution Envrionment (PXE) support

The OneConnect platform, whether it be mezzanine card or PCIe based, offers this functionality on its NIC ports, whether physical or virtual, out of the box. In addition to that functionality, where we go above and beyond is at the hardware layer we also provide TCP Offload Engine support.   With all of these technologies enabled we can still drive over 3.6m packets per second in an ethernet environment.

Moving forward with an entitlement you can easily turn on FCoE or iSCSI HBA support within the adapter as well.  The upgrade process is a simple flash update that can be applied remotely or locally on the systems using our OneCommand tools.

This introduces challenges to your I/O card vendors as they need to support it. There are two ways to do this, the right way, or, the wrong way. The right way involves the simple term HID. Human Interface Device. It’s a fancy engineering term for ‘make it easy to use’. One of the challenges about I/O devices in general is that the way they interact typically is not with easy to use/remember numbers. Whether it’s an IP address, a World Wide Name, or a hexadecimal memory range these aren’t things you easily remember.One of the largest issues when you get into the Fibre Channel world when it comes to HID issues is setting up a boot from SAN infrastructure. Outside of tools designed to automate this when you’re sitting in front of a computer you are presented with a very limited number of options when it comes to figuring out what your card’s WWN is and whether or not it can see the storage to boot from.

There are some hacks you can do where you can auto-scan and try to boot from the first thing you find. That might be fine if you’re playing around in a lab but when it comes to a production environment that’s not always a good thing. You really need to have the ability to turn on the system, search your FC network, find the device and match it to your adapter. Something a lot of us have taken for granted because every time you hit Alt-E/Ctrl-E and entered the adapter BIOS to do this.

What’s different with UEFI now is that all adapter BIOS and system BIOS are handled through a ‘main menu’ of sorts. You hit F1 on a system and now you’re presented with every ROM in the system from a main menu. The standard ‘BIOS’ things of the system are there. but now you have additional menu options that are the ROMs of the adapters.

That is, if you pay attention to the HID.

With out the HID, you get an interface that makes the Unix command line structure almost intuitive by comparison. While this may offer you ‘robust configuration options’ (aka tech speak for ‘make it impossible to do things quick & easy) at the end of the day 99% of the time you go into the firmware of a storage adapter you need to do two things. Find the storage LUN, and attach to it.

Emulex Labs will be soon posting a walk through video of how we’ve listened to our customers and made sure our UEFI integration is second to none on the BladeCenter HS22 and the 3550/3650 M2 systems that now support UEFI.

In my last post I talked about this new word everyone is tossing about the new buzz word ‘convergence’ and how I noticed that it wasn’t very converge-y (is that a word?) of them to announce a new adapter, and then only talk about one of the features that would be needed to truly ‘converge’ an adapter.

Imagine my ‘oh no they didn’t’ reaction when indeed, they announced YET ANOTHER 10Gb Ethernet Card. They even go so far to toss in a quote from someone with SAN in their name in the press release and they group the ethernet card under their ‘converged networking‘ banner on their product page but the only thing converge-y about it is that it has an ethernet port and can go in a server right next to their FCoE adapter.

Oh, and until April 30th, that chip had a Netxen brand on it. So now, I’d imagine it’s not the same device driver stack as the ethernet port on their FCoE adapter.

So there you now have two cards that ‘converge’ your network. Each with a specific function and absolutely no convergence whatsoever. So if you want 10Gb Networking & FCoE you get two adapters. Just like before. Looking at the definition of the verb convergence you get: ” (of a number of things) gradually change so as to become similar or develop something in common” or “come together from different directions so as eventually to meet”. I guess our competitor will get there eventually, but the market isn’t going to wait forever. They want a solid convergence point, and soon.

On another note, I’m going to do a couple plugs because we’re giving away some things on our website.

The first is if you download Emulex’s Converginomics Guide guide you get a chance to win a free Kindle 2. For you technical guys just download it quick and skim it. It’s fairly high level but hey a chance to win a Kindle 2 is always worth downloading a PDF file.

Also on the IBM Microsite we’re giving away a free iPod Touch via a random drawing if you answer a quick survey.

I’m happy to see some offerings in this space as it starts the conversation. However like I have previously mentioned just because you brought the horse to water, you’re not going to force them to drink. As others have learned, in the IT Industry being ‘first’ doesn’t always translate to ‘market winner’. So what does it take to be a winner?

First and foremost I strongly believe this is not a race, but a marathon. There is a sequence of things that needs to fall into place over the coming months to set the stage for FCoE technology to become to the forefront. The most important one of solidifying the standard and meeting the necessary T11 certifications.

I dislike using the way overused ‘these economic times’ because I think this would be true regardless of the financial situation but I continue to stand by my ‘not going to rip and replace‘ statement. There needs to be a compelling reason to change. I think both of our competitors missed the mark when it comes to this.

I’ve been enjoying the first five months now with Emulex and have spent the last two months touring the US and Canada and visiting with IBM’s strong Systems Engineer community that is responsible for the System X and BladeCenter portfolio and have gotten some great feedback from what their customers are telling them. I’ve also had some time to visit with some long time Emulex customers to get some feedback from them as well.

I always find it interesting in how some of the tech industry news ‘reports’ these days. When you see a quote about how ‘QLogic has a leg up on Emulex’.

Of course you see it touted over on twitter by Q as being some huge win for themselves. However in this instance the leg is more of a pirate wooden leg, and you really wouldn’t want to stand on it.

By this measurement, the ‘leg up’ is that you have an adapter that you call a CNA that can do FCOE hardware accelerated functionality with your own developed ASIC. What I find interesting is at least for now, everything they talk about this adapter is how great it is at FCoE, but no comments whatsoever about using it in conjunction with any other protocol. What good is a CNA that is positioned as an FCoE adapter and provides no other details?

CNA stands for Converged Network Adapter. The idea is you Converge your SAN data and what was formerly your network traffic onto one component. You don’t release a FCoE adapter and call it a CNA if you aren’t going to provide the same performance aspects across the ability of the card.

Back in May we demonstrated our new uCNA adapter technology at Interop Vegas. We showed TCP/IP, iSCSI, and FCoE capabilities across multiple operating systems & environments.

I’ve seen the performance claims of 250,000 IOPS per port and I’m surprised that it’s something to be proud of. Based on past claims one has to guess that this is the dreaded 512b block test that thrashes the ASIC on the card, but isn’t a real world test. The current generation 8GB/s FC adapters from Emulex can break 200k IOPS per port in this test and we’re expecting an order of magnitude improvement over this number when our OneConnect adapter.

I’d like to point out that while it’s easy for some random systems engineer employee of Emulex to say something about performance but if you compare past results where we have had significant leadership it can’t be all that hard to imagine us willing to introduce a product without being certain that we will continue our trend of 35%+ or higher performance numbers.

Right now however this is just a war of words and measuring of inconsequential numbers printed on paper. It’ll as always come down to when the rubber meets the road. Just because you got their first at the stop light, doesn’t mean you’re going to be able to keep up for the long haul.

Recently in the events we’ve held out with the IBM technical community here in the United States & Canada we’ve come across an unusual situation.

Apparently, a fibre channel adapter competitor of ours has decided it’s time to educate the world on the evil’s of heat sinks. That’s right, those innocuous little things that have been on computer components, cars, trucks, motorcycles, and any other mechanical device that generates heat for decades have secretly been trying to destroy us. They’ll starve air from your computer, make it hard to push air through, sorts of crazy claims.

Also according to them we’ve got nothing but a bag full of idiots in our engineering & development organization because we put heat sinks everywhere. They clearly know more than we do because they can make chips that don’t need heat sinks.

As a technical computer professional when you run across something this stupid you wonder what on earth they were thinking, and what kind of engineering they have that would allow them as a company to make such ridiculous claims. Especially when the company they level the claims at happens to have quality engineering people and a thermal imaging camera laying around. As the saying goes, a picture is worth 1000 words.

The ridiculous claims start with how hot we are, and we have to use a heat sink and they don’t. So what happens when you pull our heat sink off, turn on a computer and put some cards in it? I’ll let you know, because we’re doing that right now. A little sneak peak is on the same workload, it’s pretty toasty on both cards however even without a heat sink our card is cooler.

Once the data is all finished, I’ll provide the workload, the server, and the ambient input temperature for the entire test.