Home > Intel® IT Galaxy > United Kingdom > Blog > Authors > Alan Priestley

UK

5 Posts authored by: Alan Priestley

There is increased attention on the energy consumption of data centres and the amount of energy that is consumed in removing the heat generated by the IT equipment within a data centre. Upcoming legislation in the form of the UK governments Carbon Reduction Commitment will only serve to further increase the pressure on IT and facilities management to take a holistic view of their energy consumption.

One way to reduce DC energy consumption is to replace legacy server infrastructure with the latest generation power efficient servers ( more on that topic here & here). Earlier this week Intel CIO Diane Bryant took time out to talk to the press about the significant saving Intel's IT group has made by refreshing their server installed base and deploying servers based on Xeon 5500, this podcastfrom one of our financial analysts takes you thru the financial justification for maintaining a regular server refresh cycle and one of the many press reports on this topic is here

 

Server refresh is however only a part of the equation as the facilities infrastructure within the data centre also consumes power and there is a lot of debate within the industry as to whether it is possible run data centres at higher temperatures. Part of this discussion is centred around the use of external air for cooling ( free-air cooling ) as opposed to using air-conditioning units ( CRACs ) to force cool the air within the data centre.

There were reports recently that Microsoft have opened a new data centre in Dublin that uses free air cooling and also that Google's data centre in Belgium also utilises free -air cooling. BT have also been an advocate of the use of free-air cooling for some time..

For some time now the Green Grid have been evaluating the use of free-air cooling within data centres and have just made available a tool* that enables European data centre operators to easily assess the amount time they can operate their facilities with the use external air for cooling. The latest recommendations from ASHRAEas to the inlet temperatures for data centre IT equipment also facilitate the use of free-air cooling.

Intel's own IT group has also done work in this area and published a white paperthat discussed the results of running a test data centre in the Arizona desert using free-air cooling and minimal filtration on the incoming air. OK, so UK climate isn't quite as extreme as the Arizona desert ( yet ) but its still an interesting read an provides some good insight as to what the infrastructure within a DC can tolerate on the environmental front

One factor that has to be considered before taking the plunge and running your data centre with free-air cooling and at higher temperatures is that there is the risk that older IT equipment will actually consume more power as the fans within the servers will run faster and the electronics may consume more power. To counter this many OEMs are now starting to offer severs that can be safely used at extended temperatures so its worth discussing this with your equipment suppliers.

If you want more information on how to measure and improve the energy efficiency of your data centre The Green Grid are holding a Technical Forum in London on 20 October where experts from across the industry will discuss the work of the Green Grid, and the tools available to help assess your current infrastructure and plan improvements.

* The Green Grid on-line Free Air Cooling tool

0 Comments Permalink

VM 09 @ Earls Court this week brings together many of the key elements driving Cloud Computing with the software and hardwrae vendors discussing their offerings to enable IT to deploy and utilise Cloud based infrastructure.

One of the major tenets of Cloud architectures is the ability to seamlessly deploy workload anywhere within the cloud and to scale available compute resources based on workload demand. Virtualisation is the key element that enables cloud providers to deliver these capabilities to their customers and it is the availability of server hardware with in-built virtualisation support that is the underpinning of these developments.

One of the key considerations when developing virtualisation software ( the hypervisor ) is how to ensure that the guest ( i.e. virtualised ) workloads do not see that they have been virtualised, to do this it is essential that the impact of the hypervisor is minimised.

For some years now Intel has been working with the hypervisor vendors to implement hardware support for virtualisation within their processors. This support is aimed at making it easier for the hypervisor vendors to implement their code and to ensure that they are able to transparently virtualize the guest or hosted workloads.

Intel’s virtualisation technology has focused on 3 key areas – the processor, the I/O subsystem and the network interfaces.

·         Processor enhancements – new instructions and protection model that enables the hypervisor to co-exist with unmodified guest operating systems and to host multiple operating systems on the same hardware. With successive generations of Intel processors new features have been added to the processor to help the hypervisor operate and to reduce its overhead.

·         I/O subsystem – one of the limiters in early virtualisation implementations was the ability to ensure adequate I/O throughput and isolation between various workloads sharing the same physical I/O devices.  Many new Intel Xeon processor based servers now have virtualisation support built into the chipsets.

·         Network interfaces – another challenge with running multiple guest environments on the same server has been the need for them to share the same physical network interfaces and for the hypervisor to manage the separation and distribution of network traffic between the various guest virtual machines. This has been addressed with the latest network interface chips that provide hardware support to manage movement of network data directly between the LAN and virtual machines

It’s also worth noting that raw CPU performance plays a significant part in determining the number of workloads a virtualised server can host. For example if the hypervisor consumes ~10% of the available compute resource, reducing the hypervisor overhead by 10% would result in ~1% more CPU resource being available to the guest workloads, whereas increasing the CPU performance by 10% provides 10% more compute resource to the guest environments – which could equate to 1 or more addition virtual machines being hosted.

When all this is hardware support is put together with the software developments that the hypervisor vendors have been making in terms of tools to dynamically provision and move workload between various physical servers we can start to see how the underpinnings of cloud computing are being put in place.

For further reading there’s lots of good stuff in the Cloud Computing and Virtualisation tracks at the recent IDF, all the materials are here.

0 Comments Permalink

My First Cloud

Posted by Alan Priestley Oct 2, 2009

Much of what we read on the web today around cloud computing focus’s on the big US providers with their massive data centre capacity, most of which is centred around providing ‘services’ to consumers – IM, email, on-line photo galleries, backup etc. A lot of what’s written would also lead us to believe that many IT departments are going to ditch all their internal IT equipment, servers, storage etc and move wholeheartedly into the cloud.


I suspect in reality the situation is somewhat different.


For many IT organisations the first step in the direction of cloud computing will be to virtualise their existing infrastructure and workloads to provide ‘cloud-like ‘ services internally to their organisation.  Once this model has been established it then becomes viable to consider moving some non-business critical workloads out onto computer resources operated by third parties, i.e. external or public cloud services, whilst still retaining high degree of control over the application stack and data.

 

For smaller organisations without a comprehensive internal IT infrastructure the situation may be somewhat different as cloud services provides them with the option of accessing software functionality form the cloud without having to setup up and maintain themselves; these could vary from email systems, office productivity applications and all the way to complete ERP solutions.

 

So, are you planning your first cloud deployment and what will it be – an internal cloud or will you go the whole way and into the great unknown of the public cloud ??

 

If you want to find out more about your options for deploying cloud computing the IT Expo in London ( 7-8 Oct @ Earls Court ) is a good place to start, there’s also the opportunity to hear Intel experts talking about virtualisation and future compute models.

2 Comments Permalink

The Challenge is to find it !


The promise that Cloud Computing brings is that of a ubiquitous compute infrastructure that provides services accessible over the internet from anywhere, by any device, and as such Cloud Computing is currently getting lots of air-time on the web news services.

The challenge at present is much like in real life every cloud is different, offering different capabilities & services, different tools to manage, different SLA's different charging models etc - basically all clouds are unique.

Before embarking  into the great unknown of Could Computing enterprise IT managers need to fully understand what they are being offered and to look closely at their current infrastructure to determine how best to utilise Cloud Computing. Its key to understand what apps and/or workload may be suitable for moving into the cloud - for many IT departments today this starts with assessing non-mission/business critical workloads, for Intel's IT department this meant starting with our internal travel expenses tool and hosting this as a cloud service. If you haven't seen it there is a great white paper that outlines IT@Intel's cloud strategy and how they assess which applications and workload to land into the cloud.

One of the initial considerations to make when deciding to use cloud computing is whether to move directly to using an external cloud service or to utilise resources within ones own data centres to setup 'internal clouds'. For many IT departments this is a logical extension of the ongoing work to virtualize their existing infrastructure and provides a valuable learning platform before setting out into the great unknown of the external 'public' cloud infrastructure.


Some of the questions IT managers need to ask when evaluating Cloud Computing offerings are

  • What's the SLA being offered - how reliable is the service, what's the guaranteed uptime, what's the response time to fix failures ( nearly every week there are reports of cloud failures ),  what's the financial impact of loss of cloud services and how would this be recompensed.
  • What's the application latency/response time
  • How secure is the service - backup policy for data, how secure is my data, what's the isolation between my app/data and other customers in a multi-tenanted cloud.
  • Where is my app/data located - what's the impact on local data protection laws,

& the list goes on . . .

To add to the confusion around Cloud there are many different 'Service' offerings being discussed within the industry, these vary from renting raw compute resource by the hour - Virtual Machines that look like servers -  through pay-per-use for complex business applications.  Everything is a Service its just a case of what type of service you want to buy - Infrastructure as a Service ( IaaS ), Platform as a Service ( PaaS ) or Software as a Service ( SaaS ).

Today most vendors offering 'Infrastructure as a Service' have developed their own cloud platforms based around open source or commercially available hypervisors and these require specific tools and stack formatting to utilise the virtualised compute resources. Cloud vendors offering PaaS  provide application/web frameworks enabling the use of higher levels of abstraction e.g. Python, Ruby or .Net. With both IaaS and PaaS the customer needs to develop and maintain the application stack that will run in the cloud, whereas with SaaS the provider maintains the application, the customer just provides the data and pays for use of the application.

All of this further drives the uniqueness of every cloud offering. This will no doubt change over time as the major virtualisation vendors make inroads into this space and standards get developed for packaging virtual machines to be loaded into the cloud and services interfaces.

But today there is little/limited interoperability between Cloud providers and this can limit the flexibility and usability of the Cloud.

So great promises, and lots of challenges, but as the saying goes 'every cloud has a silver lining' - IT just needs to work out how to extract the benefit before the clouds all float away.

0 Comments Permalink

Green has been a buzz word in the IT industry for some time now, and there are many groups and companies focusing on helping IT understand their data centre energy consumption and methods by which this an be reduced.

 

We are now starting to see government take more interest in this as part of their policies to reduce CO2 emissions and carbon footprints. The UK government is one of the first to put in place regulation that requires companies to monitor, report and reduce their energy consumption. The Carbon Reduction Commitment requires companies to report their energy consumption and to buy allowances to offset their CO2 footprint. This scheme starts in April 2010, but takes energy consumed in 2008 as its baseline - so its too late now for a business to affect the first years reporting and they need to focus instead on how to reduce energy consumption over the coming years so as to limit the amount they need to spend on allowances.

 

Many large data centre operators and IT departments consume significant amounts of energy, both in the IT equipment and the facilities infrastructure to support the IT equipment. One of the biggest challenges today is that many IT managers do not see the electricity bills for their data centers - this usually goes to the facilities department and so they may have little or no idea of the energy costs of their infrastructure. Another factor to consider here is that many data centers are running older generation IT equipment, much of which was not designed with energy efficiency in mind and the facilities themselves were often not designed to optimize the air handling and cooling within the building or to minimize the costs of running this equipment.

 

There are a number of tools available today that can help an IT manager get a better understanding of how efficient their DC is in terms of energy usage and can help in understanding the effect of various changes on the overall energy efficiency of the data centre. One of these is the Power Utilisation Effectiveness ( PUE ) metric developed by The Green Grid which looks at the ratio of the IT load to the total facilities load within data centre. The Green Grid also publishes various white-papers and guidelines that can be used to better manage the PUE value for a data centre. There are also some good white papers produced by Intel's IT group on how to improve data centre layout and design.

 

It is also important to consider the actual IT equipment that is deployed within the data centre. Over recent years there have been many advances in server power management technology that when combined with the performance increases delivered by the latest generation CPUs can help IT reduce their overall energy consumption. With Intel Xeon 5500 based servers its possible to achieve significant levels of server consolidation, which when combined with the fact these servers typically consume less power than those they will replace can lead to significant reduction in power bills. In many cases it is cheaper to buy new servers and operate  than it is to continue running the existing installed base of servers. Intel has a simple tool, here, that can help IT make this assessment and determine how fast a payback they can achieve by refreshing their installed base of servers.

 

 

Using these tools and latest generation server technology can help an IT department make a significant contribution to lowering energy consumption and contributing to a businesses Carbon Reduction Commitment.

0 Comments Permalink