Data Center Strategies

What if Novell said YES?

In my previous post on Novell’s answer to Elliott Associates unsolicited offer, I indicated that I’d talk about Novell’s assets that many may not be considering.

For those of you following the (yet another) SCO vs. Novell UNIX copyright trial, the verdict came down in Novell’s favor yesterday (again).  Novell’s ownership of the UNIX copyrights is just one asset of many to discuss.  A comment to my blog came from Otto: “The big problem, IMHO, is that Novell owns the Unix patents. If Elliott Associates sells those to, let's say, some hyper-aggressive monopoly it could cause the Linux community still more pain.”  It’s not just a hyper-aggressive monopoly to be feared, it’s any company looking to make a buck.  This brings up the subject of patent trolls.

Elliott Associates would be interested in recouping its investments in Novell, and would sell assets to the top bidder – regardless of who that was.  Novell not only owns the UNIX copyrights, but has a number of patents that prove very important in the software industry.  It must be noted that with patents, it’s not the number, but the quality of individual patents.  It only takes one extremely good patent to inflict a sales halt on a multi-billion dollar product line if that patent were to fall into the wrong hands.  There are individuals and organizations that make their business in collecting patents from failed companies and then using those patents to extract royalties from deep pocketed vendors. These are patent trolls.

Novell has key patents in many areas, but most importantly in the areas of directory and identity services, operating system kernels, file services and systems, document management, and collaboration.  You can bet that the large software and hardware vendors in the IT market have their legal departments watching closely what is transpiring with Novell.  They would all prefer that Novell remain an independent company that they can partner with.  Savvy vendors have already struck partnerships with Novell and understand the value and protection it brings to their businesses.

For the sake of the IT industry, let’s hope that Novell stays independent, or is acquired by a friendly vendor.

[Posted by: Richard Jones]

A Busy Week (Part 1)

Apparently Intel and AMD don’t care about poor analysts trying to get their content completed by the end of the quarter, as evidenced by their decision to launch major server processor refreshes (AMD, Intel) on consecutive days this week. Needless to say there has been a vast amount of coverage from web pundits already, so I’m not going rehash it in detail.

Instead I’m going to focus on things that should matter to anybody buying servers this year. First, lets look at Intel’s Xeon 75xx and 65xx processors with respect to:

• Virtualization performance
• Memory subsystem
• System reliability
• Scale up

I’ll look at the AMD Opteron 6000 in a later blog.

Virtualization Performance

Virtualization is one of the few workloads that really make sense on these new processors, and judging by the numbers, they really do deliver. Using VMware’s VMmark benchmark, Intel claims a top score of 71.85 @ 49 tiles using 32-cores (4 x 8-core processors) on an IBM System X3850 X5. To put that in perspective, the previous best 32-core result topped out at 31.56 @ 21tiles on HP ProLiant DL785 using 8 x 4-core AMD processors. The result also trounced a 64-core result (48.23 @ 32 tiles) using 16 of the previous generation Intel XEON processors. That’s a hugely impressive result for a 4-socket system!

Memory Subsystem

This release drives the final nail into Intel’s aging “Front Side Bus” (FSB) memory architecture that dates back the mid-1990s. The FSB architecture had all the processors connected to a common memory controller (AKA “Northbridge”) so that each processor competed with its peers for access to a common pool of memory, thus becoming a bottleneck for performance. The new processors have integrated memory controllers supporting up to 16 memory DIMMs for a total of 64 DIMMs on a 4-socket server (512 GB with 8GB DIMMs), or 32 DIMMs on a 2-socket server ( 256 GB with 8 GB DIMMs).

System Reliability

In the past, x86-based servers have have used a unsophisticated approach to error handling; tell the operating system that something horrible happened and keel over in a heap on the floor. The new Xeons take a very different approach that allows the hardware and operating system to react in a much more flexible way to errors. For example if an unrecoverable memory error occurs, the operating system can look at the error and simply map that memory location out of use, or kill the process (or virtual machine) that is using the affected memory location.  In effect, Intel has blurred the line between it’s own high-end Itanium architecture and Xeon, which is good for Xeon, but not so good for Itanium.

Scale Up

The new Xeon architecture can be used to build 8-socket servers with off-the-shelf components (sometimes referred to as a “glue less design”). But they don’t stop at sockets, the Xeon 7500 family in combination with third-party developed memory hubs is designed to support configurations as large as 256 sockets (4096 threads) with 16 TB of memory! The combination of a highly scalable architecture and high system reliability has the potential to put commodity server architectures into direct competition high-end RISC architectures. For a glimpse of where the future of scale up architectures may lie, see SGI’s Altix UV announcement.

Posted by: Nik Simpson

Novell Says NO! - No Surprise

Novell officially turned down the purchase offer of $5.75/share from Elliott Associates in a letter and press release this past Saturday.  So what does this mean for the software company?

The rejection comes as no surprise.  Any board would value their company more and would want top dollar.  Note in the letter that: “Novell's Board is committed to enhancing value for Novell stockholders and believes that an exploration of alternatives is in the best interests of the Company and its stockholders.”  And that the paths they are pursuing to enhance stockholder value “include, but are not limited to, a return of capital to stockholders through a stock repurchase or cash dividend, strategic partnerships and alliances, joint ventures, a recapitalization and a sale of the Company.”

In short, Novell is saying: “Come on Elliott – or anyone else out there – you can do better than that!”  However, I think this is exactly what Elliott was hoping would happen.  Start a bidding war.  But most people would think that Novell isn’t worth that much, but that’s where they are making a mistake.  Novell has more assets and than most people realize – and most importantly, in ways that most people don’t realize.  I’ll blog about those in an upcoming blog.

Novell is open to any tactic to increase shareholder value, but they know not to jump at the first opportunity.  Most of these negotiations and possibilities will occur under the covers.  So what do I think should happen to Novell?  I’ve told Novell executives on a number of occasions that they need to find the right technology company (not investment firm) to merge with or be acquired by.  Novell continues to produce great technology, but they have never been able to market themselves out of a paper bag.  Even during the years I was there, each new chief marketing officer (CMO) came with glorious plans to improve Novell’s image.  That was the root of the problem, the Novell brand isn’t bad, it’s just become irrelevant.  The analogy I like to make to help people understand is that they need to think of how the IT world regarded Digital Equipment Corporation (DEC) in the mid 1990s.  It was known as a great technology company with excellent mini-computer systems, but the needs of the market shifted making them all but irrelevant.  The same thing has happened to the Novell brand: regarded with thoughtful awe from a past era.  The company needs to surface its technologies under a different more relevant brand today.  Note: Novell has done this with some products (eg Platespin Orchestrate was originally going to be called Zenworks Orchestrator but they latched onto a more modern relevant brand for that product) – but the crux is they need to lose the Novell name for new technologies and products.  The name needs to go down in the history books and remain revered for its past legacy.

Tell me what you think should happen to Novell?

[Posted by: Richard Jones]

Is That a Graphics Card in My Server?

For many years, one of the common factors in x86 servers has been a graphics subsystem characterized by the cheapest graphic chip the vendor could find to put on the motherboard. The logic (if you’ll forgive the pun) behind this was simple; you don’t buy servers to run graphics intensive applications, that’s what workstations are for! However, 2010 is shaping up to be different with two applications for graphics cards helping to make the case for graphics in the server.

The first of these; high performance compute clusters with graphics cards used to accelerate math calculations applications is not entirely new. NVIDIA has been pushing it’s CUDA program language graphics cards for at least 18 months with ATI/AMD expected to join the fray this year. However, the second application; accelerating graphics for virtual desktops is very new, with Microsoft’s announcements related to it’s RemoteFX protocol setting the stage for the use of high-performance graphics in servers (see here for more on RemoteFX). One of RemoteFX’s capabilities is the ability to use host-based graphics hardware acceleration to offload the graphics processing needed to support hundreds of Windows 7 virtual desktops (RemoteFX may support Vista as well, but I’d be surprised if they back-port the capability to Windows XP).

If Microsoft is successful with RemoteFX, then the real question will be how will graphics cards be integrated into today’s server platforms, and it’s not as simple as you might think. Today’s hi-end graphics cards have a number of requirements that largely rule them out for use in servers:

• Power requirements: Typical graphics cards require 150 watts or more of power, and that’s not going to be easy to satisfy in typical server designs where the power supplies tend to be small, and limited by the server form factor. It’s also quite common for the graphics card to require it’s own dedicated connection to the power supply, a feature not found on server power supplies.
• PCIe slot requirements: Graphics cards typically require a PCIe x16 slot on the motherboard, and these are rare on server motherboards where a pair of PCIe x8 slots is a more common configuration.
• Physical size: Graphics cards are big beasts, NVIDIA’s Tesla C2050 card is a full length, double width PCIe card, and that’s a lot of space to find in most servers.

So off-the-shelf graphics cards aren’t good fit for servers, and have components that aren’t needed at all such as the ability to send graphics out over VGA or DVI connections. If things are difficult for conventional rack mount servers, they are much worse for blades where the restrictions on power and physical space are even greater.

My bet is that we’ll need something a bit different if this type of hardware acceleration is going to take off. Here’s what I think we’ll see:

• A PCIe card designed for servers, with lower power requirements,a more reasonable form-factor, and a PCIe x8 (or even x4) interface.
• External, dedicated graphics engines connected via a PCIe ribbon cable to the server.  NVIDIA has already gone down this path for HPC applications with products like the Tesla S2050. This approach may work for blades as well as rack mount servers.
• Graphics cards in a blade form factor, i.e. a graphics card that takes up a blade slot and connects via PCIe over the blade chassis backplane.

The external graphics engine may also be a good place to make use of multi-root I/O virtualization that can share the graphics engine between multiple servers. Anyway this is certainly going to be an interesting space to watch as desktop virtualization becomes a mainstay of enterprise desktop strategies.

Posted by: Nik Simpson

the sleeping giant awakes - more on Microsoft's desktop virtualization announcement

Yesterday's other big news was that Microsoft’s investment in Calista is finally paying off.  Windows Server 2008 R2 SP 1 will see the introduction of Calista technologies. Now renamed RemoteFX, Microsoft has taken Calista and remade it as both a new graphics engine to power RDP and a major incentive for customers to look to Hyper-V for as the foundation for any future desktop virtualization project. 

A User-centric Microsoft Licensing Model? Not Yet, but Getting Closer

There are some physical things in life that I like to associate with as part of my identity. My Jeep Wrangler is one of them. My laptop is not. How I live and work is not defined by a physical compute device, but rather by my online identity. Email, Twitter, accessing documents and meeting notes on SharePoint, tracking client engagements in Salesforce, evaluating and testing virtualization solutions in my lab – those are all part of my day. My PC? It’s just what connects me to my online work space and personal space. Sorry Latitude D820, but you mean nothing to me.

From talks with colleagues and clients, I know I’m not alone. My device doesn’t define me, and certainly doesn’t hold all the data and applications I need to do my job. So what does all this have to do with Microsoft licensing? Well in my opinion, they’re getting it.

I’m not going to rehash much of the excellent commentary already out there on yesterday’s Microsoft announcement. See these posts for more background information on the Microsoft announcement:

• The sleeping giant awakes - Microsoft gets desktop virtualization right (Simon Bramfitt)
• Microsoft announces changes in desktop/server virtualization and VDI strategy – UPDATED (Alessandro Perilli)
• Microsoft VDI Inertia (J. Tyler Rohrer)

As Simon noted in his blog, VECD licensing is going away and the right to run a desktop OS as a server-hosted virtual desktop is now included with Microsoft’s Software Assurance (SA). For devices not covered by SA, organizations can purchase Virtual Desktop Access (VDA) licenses at a cost of $100 per device per year. Furthermore, Microsoft’s license transfer restrictions still apply. So if you want to license a virtual desktop for external contractors, you can assign a VDA license to each contractor system. If a contractor completes a project and leaves, you can re-assign the license to another contractor’s device (you just can’t reassign the license more than once per 90 days).

Simon also mentioned that “extended roaming rights” is the big deal in the announcement, and it is. While not perfect (Simon describes the issues), it’s a step toward licensing a desktop for a user and not a device (sure technically we’re still talking about device licensing, but the user can access his desktop from a myriad of personal devices). So let’s call it an alpha release of per-user licensing. Does a per-user model solve all of our problems? No. But organizations want it offered as a choice, and it’s good to see that Microsoft is listening to their customers.

Looking past the good news that came out of yesterday’s announcement, considerable work remains. Microsoft has still not addressed the service provider market. Considerable clarity is still needed for licensing virtual desktops on shared infrastructure. For example, if a user needs a Windows desktop for a week, he essentially has to pay for 90 days worth of licensing. Why? Even with VDA, the service provider technically has to associate the VDA with the subscribers physical device and can’t transfer it for another 90 days. The result is that desktop-as-a-service (DaaS) is far more costly than it should be. This problem will grow once companies like HP, IBM, and Dell offer client hypervisors, and look to offer services where user desktop VMs are automatically replicated from their personal system to the cloud. Again, this takes us back to a physical device not defining the user. For the IHVs, they get the opportunity to sell additional services to make up for the low margins they see on hardware sales. Sooner or later Microsoft will have to address this issue, and let’s hope it’s sooner.

On the support side, Microsoft’s internal application teams need to step up and offer clear support statements for the leading client and application virtualization platforms. Officially supporting App-V would be a nice first step. The push for Microsoft client applications to fully support the major client virtualization solutions must come from the top. I’m hopeful that Microsoft’s key executives will make that push.

Finally, let’s not forget that even with SA, Windows Server OS licenses cannot be virtualized (without mobility restrictions). Instead, most large enterprises have to upgrade to Data Center edition licenses (for practical purposes) for the sake of virtualizing. I talked about this issue extensively in this post, so I won’t repeat the details here. If lifting licensing constraints for client virtualization is good, I’d argue that doing the same for servers would be even better, especially if you look at the amount of servers already virtualized today.

Microsoft customers – you’re voice is being heard. Now’s a great time to pat Microsoft on the back. However, it’s not time to back off. Keep communicating your licensing needs to Microsoft. It’s clear that they are listening, and taking steps to make your life easier.

The sleeping giant awakes - Microsoft gets desktop virtualization right

Yesterday was a huge day for desktop virtualization.  Microsoft shared it's vision for the virtual desktop, corrected its past mistakes, and showed off its bright new future.

 

Right from the start Microsoft showed that it had been listening to its customers’ feedback. As of July 1st Microsoft is rolling Virtual Enterprise Centralized Desktop (VECD) into the Windows Software Assurance (Windows SA) program. This means that anyone with Software Assurance can deploy desktops locally or in the data center at no additional cost. At the same time Microsoft is extending the remote access rights so that remote isn’t tethered to a single PC in the primary users’ home. This awareness of the fact that users want flexibility around when and where they work is the key element that has been missing from Microsoft’s virtualization strategy since day one.

EMC’s global storage vision

Recently, at an analyst meeting, EMC’s president of information infrastructure products (Pat Gelsinger) gave an overview of EMC’s virtual storage vision. It’s all about the cloud and virtualization, and how both technologies carry us on a journey to a “storage from anywhere world”. As Pat spoke, nothing particularly new was in his pitch: cloud this, virtualization that, but then he drifted off into new territory….

Sometimes I wonder if perhaps I’m drinking the wrong Kool-Aid, smoking the wrong (medically approved!) ganja or just need my meds adjusted. Pat describes a global vision but lacking any detail or justification, I’m at a loss for what EMC is talking about. Sure, I get the grand scheme – make a storage container in the sky accessible to all on earth, anytime, anywhere. An aggregation of all storage into one big, dare I say it, cloud.

But at the heart of my nature I’m a practical guy. Sure, go ahead and give me all the lofty platitudes. Dream of rainbows and sunshine, but in the end I want to know where the river runs, what’s the implementation plan and how will “what-ever-it-is” make my life better? The SXSW festival – movies, music, and social interaction technologies - is in full swing here in Austin, TX – those folks know fantasy and how to spin rainbows. For the rest of us who live on terra firma, things are not so simple.

puff….

(Kudos to the cartoon author – this cartoon originated  back in the 70s) 

It would be easy to give EMC a hard time, not for their vision but their lack of practical detail. But I’d rather not go there -  it’s way too easy to throw stones at big, new ideas. I have every reason to believe that Pat and his colleagues are pretty smart and not likely smoking the ganja in their strategy meetings.

puffff

Soooo, I’ll take the high road. In fairness, Pat did point out 3 practical obstacles for his data-storage-everywhere vision – latency, bandwidth, and consistency. All these challenges revolve around the speed of light and the virtual time travel necessary to deliver a responsive storage system.

Miracles happen – engineers do it all the time.

But apart from the speed of light problem, there’s some other pressing issues that need attention, if not outright miracles. For instance, it would be great to see more work done on refining a services-based management model. No one in an enterprise setting should need to select what RAID level to use.In fact, none of the inner workings of a storage system should be exposed at all.

Some progress has been made.

IBM’s XIV, Pillar Data and Netapp (to name a few) are all on the services abstraction path for disk array management. That’s a good thing, excepting that their individual paths are not always convergent.

And, before going off to conquer the world, let’s get the local storage in a datacenter federated, aggregated and homogenized. I can hardly count how many clients I talk to with all manner of storage in their data center. Not only do they have multiple primary data systems but also a sundry of systems for backup, archive and long term storage. Heck, the vendors selling storage virtualization appliances feed on such scenarios – no one sets off from scratch with storage virtualization – they use it as a band-aid to recover from years of collecting a variety of storage platforms.

but I rant, pufff…

As Pat explained, we are all left to speculate on his vision, with some more revelations to come at EMC World. I feel like a carp being reeled in.

I can certainly speculate how a system might handle latency, bandwidth and consistency on a worldwide basis. But I won’t. EMC threw up/out the possibility – now it’s time to get practical and let us mere mortals in on the magic.

Until that day, here’s what’s most interesting about all this fantasying: the need for EMC to push its products ever further away from the (what’s becoming commodity) hardware. Except for the most extreme performance machines, storage is looking more like software packages loaded on commodity hardware. The advent of clustered commodity storage (CCS) is an early harbinger. CCS requires assembly by IT personnel, but it’s all about the software. The hardware it lies on is important to be sure, but let’s face it, hardware can be had from any number of sources.

So where is EMC trying to take us storage folk? Pat knows for sure but here’s my take:

To a land where the hardware is buried so deep it is no longer visible. Where data service, service level agreements (SLAs), and quality of service (QoS) are the lingua franca and no one needs to know what RAID stands for. And for good measure a virtual storage container distributed across the globe letting virtual machines, applications and data services live where ever it is convenient.

I like the vision, I’m worried about the details. Am I just blowing smoke?

puffff…

Comments?

Posted by Gene Ruth

Intel Takes the Wraps off Its Spring Collection

This morning Intel announced the first of its 2010 line of server processors, the Xeon 5600 family. The Xeon 5600 processors are a development of last year’s Xeon 5500, with several key differences:

• Manufactured on a 32nm process versus last years 45nm process
• Up to six cores and 12 MB of cache on each processor (50% greater than the Xeon 5500 family)
• Extended instruction set with new instructions to support AES-based cryptographic algorithms in hardware
• Reduced power consumption due to the ability to power down more unused areas of the processor

So far, the only performance numbers available are from the Intel press release and website, so we’ll have to see what the vendors announce. However, just the numbers released by Intel are pretty impressive, particularly on the VMmark benchmark where they cite a Cisco UCS B250M2 blade that reached 35.83 @ 26 tiles. To put that in perspective, compare the Xeon 5600 with it’s immediate predecessors the 5500 and 5400 families (results separated by 15 months):

CPU	VMmark	Power Envelope
Xeon X5470 3.33 GHz (4 cores/CPU)	9.15 @ 7 tiles	120 Watts
Xeon X5570 2.93 GHz (4 cores/CPU)	25.06 @ 17 tiles	95 Watts
Xeon X5680 3.33 GHz (6 cores/CPU)	35.83 @ 26 tiles	130 Watts

So in little more than a year, the performance of 2-socket Intel servers have increased by almost 4x! But it’s about more than just performance, many are now focusing on performance/watt and the improvements there are equally compelling. The following table compares estimated power costs for a sixteen server virtualization cluster built in 2008 versus a four server cluster using the latest Xeon X5680, assuming that the average power consumption is 400 Watts/server.

	Annual kW-hours	Power @ $0.1/kW-hour
16 servers @ 400 W	56,064	$5,604
4 servers @ 400 W	14,016	$1,401

Add on savings for cooling and the new servers would save in the order of $25,000 over three years just on the power bill. Add in reduced spending on hypervisor licensing etc, and the business case for replacing even relatively modern servers is pretty good. The case for replacing older servers is even more compelling. Assume each of the new servers can host virtual machine workloads equivalent to fifteen aging physical servers, so four servers can replace 60 physical servers:

	Annual kw-hours	Power @ $0.1/kW-hour
60 servers @ 400 W	210,240	$21,024
4 servers @ 400 W	14,016	$1,401

Add on savings for cooling and the new servers would save ~$120K over 3-years, more than enough to pay for themselves.

Posted by: Nik Simpson

Preempting Apple

While vendors such as Citrix and Wyse have worked hard to foster the perception that the iPhone has what it takes to be a desktop virtualization end-point, the limitations imposed by the 3.5 inch display are painfully visible. Now with the iPad just around the corner screen size is no longer a problem, but this doesn't mean that Windows on the iPad is assured of success.  Setting aside the obvious ergonomic limitations that the lack of physical keyboard imposes, perhaps the most significant shortcoming of the iPad the lack of support for multitasking. Accepting that one of the primary selling points of the iPhone/iPad is the availability of countless numbers of applications, it's doesn't take much to understand that users might want multitasking.  However, while the iPhone 3.X operating system does offer preemptive multitasking. Apple have hobbled it so that multitasking support is restricted only to those applications that Apple grants approval to (in reality this means a subset of Apple developed applications). Which leads to the rather bizarre situation where the only way to get unrestricted multitasking on an iPad is to use it as a thin-client connected back to a remote desktop.

Application developers have for the most part worked hard to overcome this limitation by exploiting the OS to the full so that application state can be remembered between sessions - something that does go a long way to ease the pain of restarting an application from scratch. However, for any application that performs remote control this is not really enough.  While the application state can be remembered the remote connection is always dropped, delaying the restart and increasing the chances that the remote session might be reset. 

Last year with the launch of the iPhone 3GS there was significant speculation that the introduction of more memory and a faster processor would see Apple loosening the shackles and allowing more applications to take advantage of the operating system's multitasking abilities but this hope was never fulfilled.  It should be no surprise then that with an iPhone 4.0 upgrade in the works for this summer stories are starting to surface that Apple will finally allow third-party apps to run in the background.

As AppleInsider explains:

From a user-facing perspective, Apple plans to deliver a multi-tasking manager that leverages interface technology already bundled with its Mac OS X operating system, according to those same people. It was requested that specifics be withheld at this time, as the iPhone Software 4.0 remains under development and reportedly has a quite 'way to go' before it's ready for prime time.

There's no information on the possibility of Apple extending this privilege to all applications, indeed given just how resource intensive some applications can be it would be foolhardy for Apple to relinquish all control especially if it compromised either the security or battery life of the iPad. Still it does provide hope for those of those who might consider an iPad as a suitable platform for accessing server hosted virtual desktops yet are frustrated by the current limitations that lack of multitasking support imposes.

If Apple fails to deliver, we can be assured that Citrix and VMware's desktop virtualization marketeers will be racing each other to announce that the they can make the iPad multitask even if Apple can't.