A tale of two CPU architectures

Posted on By Mark Wilson

AMD Opteron Intel XeonLast week I wrote about the VMware infrastructure that I’m trying to put in place. I mentioned that my testing has been based on HP ProLiant DL585 servers – each of these is equipped with four dual-core AMD Opteron 8xx CPUs and a stack of memory. Half of the initial infrastructure will use new DL585s and the intention is that implementing these servers will release some recently-purchased HP ProLiant DL580G3s for an expansion of the infrastructure. Because the DL580G3 uses an Intel Xeon MP (formerly codenamed Paxville MP) processor, the difference in processor families requires a separation of the servers into two resource pools; however that’s not the real issue. My problem is justifying to an organisation that until now has exclusively used Intel processors that AMD units provide (as my CTO puts it) “more bang for our buck”.

The trouble is that the press is full of reports on how the new Intel Xeon 51xx CPUs (formerly codenamed Woodcrest) out-perform AMD Opterons, where AMD has been in the lead until now; but that’s in the 2-processor server space and I’m not hearing much about 4-way units.

All of this may change tomorrow as, at today’s VMware Beyond Boundaries virtualisation roadshow, Richard Curran, Director of Marketing for the Intel (EMEA) Digital Enterprise Group, hinted about an impending announcement; however an HP representative expressed a view that any new CPU will just be to plug the gap for a few months – the real performance boost will come in a few months time with the next generation of dual-core multiprocessor chips (in the same way that the Xeon 50xx chips, formerly codenamed Dempsey, preceded the 51xx Woodcrest).

Leaving aside any other server vendors, I need some direction as to which 4-way server to buy from HP. HP ProLiant DL580G3s would allow me to standardise but the newer HP ProLiant DL580G4s are more powerful – using the Xeon 71xx chips (formerly codenamed Tulsa) with Intel VT virtualisation support – and, based on list price, are significantly less expensive. Meanwhile, HP’s website claims that ProLiant DL585s are “the best performing x86 4-processor server in the industry” and they cost slightly less than a comparably-specified DL580G4 (again, based on list price), even before taking into account their lower power consumption.

Speaking to Intel, they (somewhat arrogantly) disregarded any reason why I should chose AMD; however AMD were more rational, explaining that regardless of the latest Intel benchmarks, an Opteron is technologically superior for a two main reasons: the hypertransport connection between processor cores; and the integrated memory controller (cf. Intel’s approach of using large volumes of level 3 cache), although the current generation of Opterons only use DDR RAM. Crucially though, AMD’s next-generation dual-core Opterons are socket-compatible with the forthcoming quad-core CPUs (socket F) and are in the same thermal envelope – allowing for processor upgrades – as well as using DDR2 memory and providing AMD-V virtualisation support (but in any case I’ll need to wait a few months for the HP ProLiant DL585G2 before I can buy a socket F-based Opteron 8xxx rack server from HP).

As my virtualisation platform is based on VMware products, I asked VMware which processor architecture they have found to be most performant (especially as the Opteron 8xx does not provide hardware support for virtualisation; although there are doubts about whether ESX Server 3.0 is ready to use such technology – I have read some reports that there will be an upgrade later). Unsurprisingly, VMware are sitting on the fence and will not favour one processor vendor over another (both AMD and Intel are valued business partners for VMware); of course, such comparisons would be subjective anyway but I need to know that I’m making the right purchasing decision. So I asked HP. Again, no-one will give me a written opinion but two HP representatives have expressed similar views verbally – AMD is still in the lead for 4-way servers, at least for the next few months.

There are other considerations too – DL580s feature redundant RAM (after power and disk, memory is the next most likely component to fail and whilst ECC can guard against single-bit failures, double-bit failures are harder to manage); however because the memory controller is integrated in each CPU for an AMD Opteron, there is no redundant RAM for a DL585.

Another consideration is the application load – even virtualised CPUs are perform differently under different workloads: for heavily cached applications (e.g. Microsoft SQL Server or SAP), an Intel architecture may provide the best performance; meanwhile CPU and memory-intensive tasks (e.g. Microsoft Exchange) are more suited to an AMD architecture.

So it seems that it really is “horses for courses” – maybe a split resource pool is the answer with one pool for heavily cached applications and another for CPU and memory-intensive applications. What I really hope is that I don’t regret the decision to follow the AMD path in a few months time… they used to say that “nobody ever got fired for buying IBM“. These days it seems to be the same story for buying Intel.

UK white pages available online

Posted on By Mark Wilson

I don’t know how long this has been available but I just noticed a link on the BT website to the Phonebook – the UK’s residential and business white pages service. Many years back I used to have some sort of dial-up terminal access to a directory lookup system but I didn’t know it was now available on the Internet. Won’t find me though… I’m ex-directory!

VMware Beyond Boundaries virtualisation roadshow

Posted on By Mark Wilson

VMware Beyond Boundaries

It’s conference season and I’ll be missing the European Microsoft TechEd IT Forum this year for two reasons: firstly, it clashes with my son’s birthday; secondly, it’s in Barcelona, and last time I attended a TechEd there I found it to be less-well organised than conferences in other European cities (e.g. Amsterdam). There’s also a third reason – it’s unlikely I’ll get budget approval to attend – but the reasons I’ve already given mean I won’t even be trying!

Given my current work commitments, one conference for which I should find it reasonably easy to justify is VMware’s VMworld; however, nice though a trip to Los Angeles might be (actually, there are other American cities that I’d rather visit), I’m expecting to have gained a new son or daughter a few weeks previously and leaving my wife home alone with a toddler and a newborn baby for a week might not be considered a very good idea. With that in mind I was glad to attend the VMware Beyond Boundaries virtualisation roadshow today at London’s Excel conference centre – VMware’s first UK symposium with 500 attendees and 27 sponsors – a sort of mini-VMworld.

Whilst I was a little annoyed at having arrived in time for the first session at 09:30 and VMware apparently being in no hurry to kick off the proceedings, it was a worthwhile day, with presentations on trends in virtualisation and increasing efficiency through virtualisation; live demos of VMware Infrastructure 3; panel discussions on realising business benefits from a virtualisation strategy and recognising when and how to virtualise; real life virtualisation experiences from BT and Nationwide; and a trade show with opportunities to meet with a number of hardware vendors and ISVs.

I’ll post some more about the most interesting sessions, but what follows is a summary of the key messages from the event.

One feature in the introduction session was a video with a bunch of children describing what they thought virtualisation might mean. Two of the quotes that I found particularly interesting were “virtual kind of means to me that you put on a helmet and then you’re in a different world” and the idea that I might “use it to get money and do [my] homework”. Actually, neither of those quotes are too far from the truth.

Taking the first quote, virtualisation is a different world – it’s a paradigm shift from the distributed server operations model that we’re used to in the “WIntel” space – maybe not so radical for those from a mid-range or mainframe server background, but a new style of operations for many support teams. As for the second quote – it is possible to save money through server consolidation which leads to savings in hardware expenditure as well as reduced power and heat requirements (one CPU at a higher utilisation uses less power than several lightly-loaded units) and consolidation (through virtualsiation) also allows organisations to unlock the potential in underutilised servers and get their “homework” done.

Indeed, according to IBM‘s Tikiri Wanduragala, server consolidation is the driver behind most virtualisation projects as organisations try to get more out of their hardware investment, making the most of computing “horsepower” and looking at metrics such as servers per square inch or servers per operator. Realising cost savings is the justification for the consolidation exercise and virtualisation is the enabling technology but as IDC‘s Thomas Meyer commented, he doubted that a conference room would have been filled had the event be billed as a server consolidation event, rather than a server virtualisation one. Wanduragala highlighted other benefits too – virtualsiation is introducing standard by the back door as organisations fight to minimise differences between servers, automate operations and ultimately reduce cost.

Interestingly, for a spokesman from a company whose current marketing message seems to be all about high performance and who is due to launch a (faster) new chip for 4-way servers tomorrow, Intel‘s Richard Curran says that performance per Watt is not the single issue here – organisations also want reliability, and additional features and functionality (e.g. the ability to shut down parts of a server that are not in use), whilst Dell‘s Jeffrey Wartgow points out that virtualisation is more than just a product – it’s a new architecture that impacts on many areas of business. It also brings new problems – like virtual server proliferation – and so new IT policy requirements.

Somewhat predictably for an organisation that has been around since the early days of computing, IBM’s response is that the reactive style of managing management console alerts for PC servers has to be replaced with predictive systems management, more akin to that used in managing mid-range and mainframe servers.

Of course, not every organisation is ready to embrace virtualisation (although IDC claim that 2006 is the year or virtualisation, with 2.1 million virtual servers being deployed, compared with 7 million physical server shipments; and 46% of Global 2000 companies are deploying virtualisation technologies [Forrester Research]). Intel cited the following issues to be resolved in pushing virtualisation projects through:

  • Internal politics, with departments claiming IT real estate (“my application”, “my server”, “my storage”).
  • Skills – getting up to speed with new technologies and new methods (e.g. operations teams that are wedded to spreadsheets of server configuration information find it difficult to cope with dynamically shifting resources as virtual servers are automatically moved to alternative hosts).
  • Justifying indirect cost savings and expressing a total cost of ownership figure.

IDC’s figures back this up with the most significant hurdles in their research being:

  • Institutional resistance (25%).
  • Cost (17%).
  • Lack of technical experience (16%).

The internal politics/institutional resistance issue is one of the most significant barriers to virtualisation deployment. As Tikiri Wanduragala highlighted, often the line of business units hold their own budgets and want to see “their machine” – the answer is to generate new business charging models that reflect the reduced costs in operating a virtual infrastructure. Intel see this as being reflected in the boardroom, where IT Directors are viewed with suspicion as they ask for infrastructure budgets – the answer is the delivery of IT as a service – virtualisation is one shared service infrastructure that can support that model, as Thomas Meyer tagged it, a service oriented infrastructure to work hand in hand with a service oriented architecture.

For many organisations, virtualisation is fast becoming the preferred approach for server deployment, with physical servers being reserved for applications and hardware that are not suited to a virtual platform. On the desktop, virtualisation is taking off more slowly as users have an emotional attachment to their device. HP‘s Iain Stephen noted that there are two main technologies to assist with regaining control of the desktop – the first is client blades (although he did concede that the technology probably hit the market two years too late) and the second is virtual desktops. Fujitsu-Siemens Computers‘ Christophe Lindemann added that client blades simply take the desktop off the desk is not enough – the same management issues remain – and that although many organisations have implemented thin client (terminal server) technology, that too has its limitations.

Microsoft’s dynamic systems initiative, HP’s adaptive infrastructure, Dell’s scalable enterprise, IBM’s autonomic computing, Fujitsu-Siemens Computers’ dynamic data centre and IDC’s dynamic IT are all effectively about the same thing – as HP put it “[to] deliver an integrated architecture that helps you move from high-cost IT islands to lower-cost shared IT assets”. No longer confined to test and development environments, virtualsiation is a key enabler for the vision of providing a shared-service infrastructure. According to IDC, 50% of virtual machines are running production-level applications, including business-critical workloads; and 45% of all planned deployments are seen as virtualisation candidates. It’s not just Windows servers that are being virtualised – Linux and Unix servers can be virtualised too – and ISV support is improving – VMware’s Raghu Raghuram claims that 70% of the top ISVs support software deployed in a virtual environment.

Looking at server trends, the majority of servers (62.4%) are have a rack-mount form factor with a significant proportion (26.7%) being shipped as blades and pedestal/tower servers being very much in the minority [source: IDC]. Most servers procured for virtualisation are 2- or 4-way boxes [source: IDC] (although not specifically mentioned, it should also be noted that the VMware licensing model, which works on the basis of pairs of physical processors, lends itself well to dual-core and the forthcoming multi-core processors).

Virtualisation changes business models – quoting Meyer “it is a disruptive technology in a positive sense” – requiring a new approach to capacity planning and a rethink around the allocation of infrastructure and other IT costs; however it is also a great vehicle to increase operational efficiencies, passing innovation back to business units, allowing customers to meet emerging compliance rules and to meet business continuity requirements whilst increasing hardware utilisation.

Summarising the day, VMware’s Regional Director for the UK and Ireland, Chris Hammans, highlighted that virtual infrastructure is rapidly being adopted, the IT industry is supporting virtual infrastructure and the dynamic data centre is here today.