Here is how I see the rack vs. blades debate: No solution is right all the time! There are situations in which racks are better, and there are situations where blades are better.
Before I go into the use cases, let me clarify one or two points for both architectures:
Racks are better under the following conditions:
Blades are better under the following conditions:
Blades have come a long way since the early days of very few options and limited expandability. Most early blade servers only had one or two NICs, limited storage, no Fibre Channel support, and limited CPU and memory, which made them poor choices for virtual hosts. That’s all changed in recent years as blade technology has evolved and no longer has the limitations of earlier blades, making them ideal for virtual host servers. Modern blade servers can support up to 16 NICs, four quad-core processors and multiple Fibre Channel or iSCSI HBA adapters. When considering blade servers in your environment as an alternative to traditional rack mount servers, you need to know the advantages and disadvantages of each and why you might choose one type over another.
Some reasons you might choose blade servers over traditional servers:
Some reasons you might choose traditional servers over blade servers:
Many people that use blade servers as virtual hosts often take advantage of the boot-from-SAN feature so they don’t need internal storage on their blade servers. The choice between blade and traditional servers often comes down to personal preference and what type of server is already in use in your data center. Some people like blades, others don’t. Regardless of which server type you choose, they both work equally well as virtual hosts.]]>
Last week, VMware and HP announced that at the end of March, VMware ESX 3i will be packaged on 10 models of HP ProLiant servers. So do embedded hypervisors like ESX 3i represent the next stage of the virtualization evolution?
Of course VMware seems to think so, saying the integrated offering will provide “greater speed and simplicity for customers new to virtualization, as well as increased capacity expansion for customers who already use VMware’s data center virtualization and management suite, VMware Infrastructure 3 (VI3).”
Will this optimism translate into increased virtualization in the enterprise? VMware and virtualization expert Andrew Kutz thinks that the exclusivity of the plug-and-play capability of 3i on HP is a stretch:
Plug-and-play is another no-win for 3i. The plug-and-play functionality of 3i is as artificial as its simplified management. VMware asserts that independent hardware vendors (IHVs) will be able to ship servers with 3i directly to the customer, where the customer can simply plug the box into the network and storage, boot it, and presto: installation complete. That’s fantastic! But I can order a server from an IHV with ESX 3 pre-installed on it today. The difference is that VMware has added this data center plug-and-play functionality exclusively to its 3i product. There is no reason that it cannot work with 3.0 or 3.5 as well. This is just another example of a company trying to promote a new product with features that do not have to be exclusive; they are exclusive only because someone decided they should be.
While Kutz believes that 3i is a significant step up, he says on SearchVMware.com that “ESX 3i is simply an evolution, not a revolution.”
The biggest change between ESX 3i and its predecessors (ESX 3.0, 3.5) is that with 3i, agents cannot be installed on a host. Erik Josowitz, the vice president of product strategy at Austin, Texas-based Surgient Inc., a virtual lab management company, says that for independent software vendors, “VMware’s roadmap for virtualization management runs through VirtualCenter.” Putting 3i on solid state “sends a clear signal that VMware doesn’t want people installing on the host anymore,” according to Josowitz. He notes that “from a security standpoint, it’s a good thing,” since it locks down the partition that used to be available under the host, thus keeping out any applications that might weaken a system. But now, organizations that want to work with blended images will need to architect their tech support to talk through VirtualCenter rather than a host agent.
While the solid-state product promises plug-and-play deployment of VMware’s thin hypervisor product on HP’s ProLiant servers, some analysts are still saying, “Don’t believe the hype about 3i.” Citing problems with monitoring and scaling of 3i, the ToutVirtual blog complains that 3i is “a complete disappointment” at general release. “Combine this weak infrastructure design issue with the fact that you can not get any realistic information out of the hardware state of a 3i server,” makes VMware ESX 3i “dead on arrival.”
But SearchServerVirtualization.com expert Rick Vanover begs to differ. Vanover holds ProLiant servers in high esteem, and if ESX 3i is good enough for HP, then it’s good enough for him:
I’ve worked on many different types of servers, and I think the ProLiant servers are superior. The big reason is that the ProLiant BL blade series do not have a competitor to the Insight Control Environment. Further, the Integrated Lights-Out 2 Standard Blade Edition (or iLO) is a better management interface compared to its competition. If VMware takes HP as a partner (or at least as their first partner) for an ESX 3i supported platform, I would choose it in a heartbeat.
But does it really matter that 3i is overhyped? Major vendors now put 3i inside their servers. This reduces the need for major evaluation and opens the door for IT shops to choose servers with “3i inside” and use it when and how they want.
What do you think? Leave us a comment below or send us your thoughts.]]>
Processor and memory inventory
The newest blade servers can run 4 sockets and 4 cores in one blade, and one model in particular that was favorably discussed is the HP ProLiant BL680c series. This is great for virtualization implementations with an incredibly small footprint. With the BL680c, each blade can house up to 128 GB of RAM. ESX 3 and Microsoft Windows Server 2008 are supported operating systems for virtualization implementations for this series of blades. One important note on the HP blade series is the Virtual Connect product for network connectivity. Fellow TechTarget contributor Scott Lowe covers this well in a recent tip.
You have to love the small footprint
With the momentum of virtualization migrations not slowing, the small footprint is very welcome in crowded data centers. The BL680c can have 80 hosts of the speck above in one 40U rack with four enclosures! Using general purpose servers would take at least double the space to get the same number of virtual hosts.
Given the very small footprint of the blade server, there are some limitations to connectivity. While the BL680c excels in most areas, it is limited to only three expansion interfaces for additional networking and fiber channel connectivity. Most implementations, however will be able to meet their connectivity requirements from the available options.
A smaller issue may be power sources. Blade servers will generally take different power sources compared to standard general purpose servers. The trade off is that in feeding a blade server a L15-30P outlet you may not need a power distribution unit (PDU). The PDU may take the same L15-30P interface, so some planning on your power sources and availability to get the correct sources available.
The current generation of blade servers are serious contenders for virtualization hosts. The small footprint only makes the case more compelling. As the blades now are able to offer comparable performance specs of the traditional server counterparts, we should consider them for the host hardware environment.]]>
With my box racked up and plugged in, I grabbed the ISO for CentOS 5.1 and gave it a whirl. On a side note, it installed perfectly on a Dell PowerEdge 1950 w/ SAS disks in a RAID 5 array on a PERC5 card. In spite of some people having problems related to the RAID setup, mine went through flawlessly. (Apparently there is a known issue with multiple arrays on a single card and GRUB’s placement on the wrong array.)
After getting the OS up and running, I gave it a whirl using a FAQ I found at Nixcraft to install VMware server.
I did it once on a fully-updated install, complete with the updated kernel packages, and it bombed out. Going back, and using an updated kernel worked flawlessly. To sum up the process, here’s what to do:
# rpm -ivh VMware-server-<version>.rpm
yum install libXtst-devel libXrender-devel
# yum install xinetd
“What about hardware decisions — should data center managers be considering scale-up instead of scale-out?”
My response was:
“I personally prefer a scaled-up approach because there is a reduction in ongoing costs, such as power, space, cooling, and physical maintenance. Also, the complexity factor is reduced when there is less hardware to manage. An exception to that would be data centers without existing centralized storage — the initial acquisition becomes more expensive in scale-up operations if a SAN infrastructure is not already in place.”
I’m guilty of being one of those people that says “Durnit, why didn’t I say this or that?” or “Dangit, why didn’t I quantify that a little more?” even well after the fact, making me perhaps my own worst critic. In this case, I really felt I left some stuff unsaid. One item that irks me about that answer is that I should have made more mention of blades. I hate blades in their current incarnation. I think they’re the worst idea in IT – they’re hot, cramped, delicate, with slower components and limited expansion ports – if you name something about a blade, I can find a reason to hate it. That said, I shouldn’t have left them out of my line of thought – a good IT Manager needs to consider uncomfortable things, difficult things, even distasteful things, when looking at something impactful. Or so says the wisdom of Frank Hayes, to whose articles I often find myself nodding to the affirmative while reading. So, here goes.
Blades are hot – they have limited cooling options built-in. That’s often a “value-add” (choke) of specialized rack systems and chassis systems provided by third-party vendors. Here’s a few links to illustrate the point:
A rack of big-honkin’ boxes will make you feel toasty on the parts next to their fans. A rack of blades will cook you medium-well given enough time. To prevent the data equivalent of multiple mini-supernovas you need to install the correct cooling – the correct tonnage of AC, hot and cold rack aisles, proper ventilation, air temperature monitors, system heat monitors, etc. In many data centers, the cost of new construction (or re-construction) may very well exceed even long-term cost savings from server consolidation, and even if you can afford the construction and still come out with positive ROI, that cooling comes at a monthly utility cost – you must increase your power consumption to keep things cool.
That said, this is where virtualization has been proven out over the last decade as a way decrease the number of servers and offload them to blades. That may mean that you can remove enough servers to use your existing heat management systems in a more focussed way and not have to break the bank. Even if it’s a five-to-one ratio of servers removed to virtualization-equipped blades added, you’re coming out ahead. Add in centralized storage systems to connect to the blades and the scales may well tip back in favor of Mr. Heat Miser again, but probably not. Getting a ten-to-one ratio means blades are a winner. This is assuming a large server consolidation via virtualization project. If it’s not a big percentage of your boxes being affected, you’ll be back in the hot seat, quite literally.
Ever need five or more NICs for a virtualization host? I have. If I had blades, I’d be using three blades to get that done, assuming dual nics, and five or more on single-nic blades. That means more blades, more virtualization software licenses I don’t need, more hardware to fail, and more physical boxes when what I want to do is REDUCE the number of physical boxes. Right now server blades are still too young – many vendor’s products have all the components are included on the blade, and not modular enough. PC blade systems have it a little better – some limited peripheral connectivity at the user-site (see this link for one manufacturer’s solution), but still, it’s an entire box in a chassis with all the difficulties of expanding that micro-sized PCs and laptops have.
So, I think it’s safe to say that I still hate traditional blades. But I think they’ll be the saviour of the data center soon, and then I will love them. Why? Because here’s my ideal blade system: a truly modular system that will change everything about blades. The best part, it’s available now from several of the larger vendors. The changes are part of a new design “paradigm” (please note my bias against that word) – the end-result is a blade system where the blades can be NICs or other devices, as needed and plugged into the chassis, connected in either a physical layer with ye olde jumper or a software layer (in the chassis management software, perhaps). Lets say I get a blade and I need to put ESX on it, but I need six NICs because of guest system network i/o requirements… ok, I get another blade with a quad-NIC on it, plug it into the chassis, and configure it – voila, a single computer with five or six NICs in two blade slots, using one license. Or perhaps I need ten USB connectors for some virtualized CAD desktops, which require USB key fobs in order to use the CAD software – I plug in a server blade and a USB blade, configure it, and voila, one server, ten USB ports, one license. Expand that out far enough, and you can have whatever you need in terms of peripherals in a blade chassis. If you go to IBM’s website, you get a whole panopoly of choices – switchblades (that one always give me a chuckle) and NIC blades are readily available for expanding your blade chassis out to do more than just host some servers. HP upstages them a bit and has a great product out now that provides PCI-X and PCI-e ports. This is from their website:
“Provides PCI-X or PCI-e expansion slots for c-Class blade server in an adjacent enclosure bay.
- Each PCI Expansion Blade can hold one or two PCI-X cards( 3.3V or universal) ; or one or two PCI-e cards(x1, x4, or x8)
- Installed PCI-X cards must use less than 25 watts per card. Installed PCIe cards must use less than 75 watts per PCIe slot, or a single PCIe card can use up to 150 watts, with a special power connector enabled on the PCI Expansion blade.
- Supports typical third-party (non-HP) PCI cards, such as SSL or XML accelerator cards, VOIP cards, special purpose telecommunications cards, and some graphic acceleration cards.”
This is interesting – a couple of PCI-e quad-NICs in one of an expansion unit and my NIC requirements are set. Or perhaps a couple of PCI-e USB add-in cards. Or a high-end PCI-X or PCI-e video card. Ok that gets troublesome when you need a lot of them – you can wind up with one blade and a chassis full of expansion slits containing video cards – the cost might not be worth it.
In any case, this dramatically changes my view on scaling up or out. Right now, I still stand for scaling up because blades don’t work in my enviornment – I have heat problems. I have space problems too, which blades could solve, but not with my heat problems. I prefer to buy larger-sized servers with lots of expandability (DL300 and 500 series, PowerEdge 2000 and 6000 series, etc.) and add in NICs as needed rather than buy blades or 1U boxes because I can do more with these larger-sized machines even though they take up more room. I fully expect that to change in the future – at some point I see myself stopping with the scaling up and starting with the scaling out – only I expect the “out” part of that will involve a lot less real estate and more options than currently available.]]>
[kml_flashembed movie="http://www.youtube.com/v/H5I5TAGZ-1w" width="425" height="350" wmode="transparent" /]
Virtual Iron's Mike Grandinetti provides insights about the synergies between virtualization, blades, server consolidation and iSCSI in this interview with Jan Stafford, SearchServerVirtualization.com's senior site editor.
After sharing his opinions on the pros and cons of the three main styles of server virtualization, Golden sounded off on the most-commonly-used hardware platforms for server virtualization. Here’s a summary of his analysis:
Server type: x86 32-bit
Example: Dell PowerEdge
Applications: Client virtualization; test and development environment
Pros: Widely available; inexpensive; IT skills widely available
Cons: Memory limitation; poor virtualization scalability
Golden says: “Repurposed machines save money in the short term, but they don’t scale very well. You need more robust memory, in particular.”
Server type: x86 64-bit
Example: HP BladeSystem
Applications: Client virtualization; midrange-to-large server virtualization
Pros: Powerful; similar skills to x86 32-bit; larger memory possible
Cons: May be limited in scalability depending upon machine design
Golden says: “64-bit blades are very powerful and offer high density, but they do pose power and cooling challenges.”
Server type: x86 64-bit specialized hardware
Examples: Sun SunFire; IBM System X
Applications: Large server virtualization deployments
Pros: Designed for high-performance scalability; large memory support
Cons: New hardware type for operations personnel; can be costly
Golden says: “This class of server offers the optimal virtualization platform for large-scale virtualization deployments, but their prices may be prohibitive for most organizations. You also have to figure in the cost of training your IT staff into the equation.”
Got questions about servers for virtualization? Disagree with Bernard’s assessments or have something to add? Bernard is a resident expert on SearchServerVirtualization.com and is available to respond to you. Please comment below or write to me at firstname.lastname@example.org.]]>
SSV: How big a deterrent to buying blade servers is power and cooling, based on your observations at the Summit? What cool things are being done about it?
Goldworm: Power and cooling and space are issues for most users, even in trying to expand their rack-n-stacks. Many of them were there because they know they have to do SOMETHING, because they can’t go on like they are. Often there is a list of easy (and not expensive) steps which can be taken, before going to more drastic measures (like liquid cooling). Planning help is available from folks like Eaton and APC, as well as HP and IBM, and others. Advances in hardware and software are continuing to come, with smarter power management, shutting down unneeded processors based on utilization, etc. Processing power per watt is continuing to improve.
SSV: Were virtual desktops — via appliance virtualization, VDI (virtual desktop infrastructure) and other models — hotter than you thought, in terms of interest?
We expected virtual desktops to be a hot topic and it was. As people get more comfortable with server virtualization, and start looking at Vista on desktops, virtualization for the desktop and applications are becoming a serious topic. I view this area as a continuum, with different approaches offering benefits for different use cases (from VDI to Citrix to the new IBM workstation blade). I think we’re hitting the tip of the iceberg here.
It’s hot and users are struggling to understand how it all fits together.
SSV: Looking back at the Summit, what are your overall impressions about the state of blades and virtualization after the Summit?
Goldworm: People have been hearing more about blades for the past year or two, often with a lot of warnings. Many came to the summit looking to get a better understanding of the benefits and the “gotcha’s” and were pleasantly surprised with the progress made in the past year, particularly relative to virtualization. Many of the customers we spoke with were very excited about the benefits that blades and virtualization could bring them, and many seemed to be hearing up-to-date information for the first time (including from their own vendors like IBM, HP and VMware).
As users and channel partners are getting more educated, we will see more and more of the marriage between blades and virtualization.]]>
Newell has more field experience working with blades than anyone else I met at the summit. As U.S. Client Services Manager for Halian, Inc., — a U.K.-based global IT services organization –he has worked on blades implementations in banking, pharmaceutical, government and other types of businesses.
The top 5 reasons to use blade servers:
1. They’re tiny. Blades conserve data center floor space better than any other server option. If your floor space is at a premium, then check out blades.
2. They’re dense. Combined with virtualization, blades give you the most compute power per square inch of any server.
3. They’re easy to deploy. Today’s blade server toolsets allow for ease of server deployments. The cabling, power and much more are built into the chassis, so there’s less to do when you slip the box in its slot. Virtualize and speed of deployment increases more.
4. They’re a good fit for lab environments. “Blades and virtual servers provide great architectures for lab, testing, and development environments,” Newell said.
5. There will be no more snakes on your plane! Those cables roping around your data center will disappear, as blades have far fewer power and network cables.
Put all these uses and benefits together. Mix well. Then, watch TCO get TKOed. Typically, corporate processes significantly increase overall server deployment time, leaving you with lower overall total cost of ownership, even though upfront costs may be higher.
Here’s the big if, and and but:
“Power and cooling concerns are real! The power consumption/square foot in a blade-based data center are significant…like 25,000 watts per chassis.”
So, do your homework, and evaluate cooling requirements and power consumption as a part of your overall cost for hardware deployment.
“Returns take numerous years due to the significant capital required within a data center environment,” Newell said. “Smaller environments may see faster returns.”
In other words, good things come to those who plan, deploy and wait.
Want more info on reasons to or not to use blades? Check out these links:
Why wed blade servers to virtualization?; Barb Goldworm’s guide to blades and virtualization; Former Morgan Stanley exec praises blades; and Blade servers dominate market by 2009.