Weekly Highlights – May 18, 2012

Posted on May 18, 2012 by GlassHouse

The Current Data Center Design and Implementation Shift
Computerworld – May 15, 2012

This article discusses the shift many public and private entities have made from large hardware and software purchases in the form of components of servers, storage and applications to pre-packaged bundles, building blocks and/or PODS. This trend is part of a more modular data center design approach that is based on vendor tested and certified reference architectures. Such reference architectures include hardware and software components that have been selected, tested and tuned by vendors instead of IT staff. Reference architecture can have a dramatic impact on IT efficiency, and it will simplify building out data center infrastructure.

Computing and Mobile Processing Saves Small Business
Forbes – May 16, 2012

This article discusses the rising popularity of cloud computing and mobile devices as a means to prepare small businesses for disaster. The cloud is being increasingly used for offsite storage and backup and allows for an easier disaster recovery process. Whereas a physical server may be destroyed, the cloud server will still be intact. In addition to the rising popularity of cloud computing, mobile processing is gaining traction in this same regard. This gives business owners the ability to process payments through smartphone or mobile devices, which helps to save a business in case of a disaster. If a physical store and/or company software has been damaged, a business owner can still take payments with their mobile card readers.

Forget Offshore Tax-havens & Meet the Offshore Cloud
GigaOM – May 16, 2012

This article discusses how Calligo has started leasing servers at data centers on the Jersey and Guernsey Islands. This strategy provides cloud options for companies that need to keep their data offshore. Calligo uses Nicira’s software to create software-defined networks inside and between its data centers to create a seamless and unified cloud for clients while also ensuring that specific security and data routing protocols are kept in place.

Why Facebook’s IPO is Good for the Data Center Industry
Data Center Knowledge – May 18, 2012

This article discusses Facebook going live today on the NASDAQ exchange and the positive effect this will have on the data center industry. Facebook’s existence and growth relies on Internet infrastructure and it has already invested more than $1 billion in the infrastructure that powers its social network, which now serves more than 900 million users around the globe. The company spent $606 million on servers, storage, network gear and data centers in 2011, and expects to spend another $500 million this year. That spending has boosted data center construction, jobs and innovation in the hardware ecosystem.

Posted in Cloud Computing, Data Center, Disaster Recovery, Infrastructure Operations, Virtualization | Leave a comment

Behind the Technology Curtain

Posted on May 17, 2012 by Jay Seaton

There is something strangely incongruous about the world of IT infrastructure. Trillions are spent on new technology every year, while IT budgets seem to be perpetually in a state of being cut.  In fact Gartner predicts that more than $1.1 trillion will be spent this year on IT hardware, software and equipment, with more than three times that ($3.75 trillion) spent on IT when services are added in, up 2.5% over 2011 (http://www.gartner.com/it/page.jsp?id=1888514).

Much like the new car industry, a significant percentage of that technology spend may be driven less by real need than by creative rationalization.  Often what we are witnessing is the next level of the BS-NOS (bright, shiny new object syndrome).

And the reality is that the technology costs are just the tip of the iceberg.  A technology purchase usually follows months of planning, assessments, evaluations and a bid process.   The new technology typically requires new skill sets that need to be recruited, hired and trained.  It rarely is fully compatible with existing systems.  In-place systems cannot simply be replaced.  Workarounds and programs to interface new and existing systems must be developed, tested and deployed.  The many months of ‘pre-purchase’ work are followed by many months of ‘post-purchase’ deployment, refinement and training.   By the time this cycle is completed, the “new” technology frequently no longer matches the business requirements.

Yet the technology juggernaut rolls on.

Data storage and protection, particularly backup and disaster recovery, has undergone major technology changes.  A decade ago, server virtualization was not much more than a research project.  Today, virtualization technology has revolutionized data center infrastructure design, with server, desktop and application virtualization implementations commonplace.   Cloud technologies are flooding the market to help drive (capitalize on) the world of IT to potentially its most dramatic transformation yet, promising the realization of a true services-based model for IT.

But behind the technology curtain, the fundamental tenet that a successful IT environment has to be based on more than just good technology remains.  A highly productive and cost-effective IT capability requires a service model that includes clearly defined standards and service, effective metrics for both cost and service delivery, consistent operational delivery and organizations that are aligned.   All of these elements are essential.

What are far too common are IT organizations that embrace new technologies but somehow never fully realize the anticipated benefits.  Costs continue rising faster than expected.  User satisfaction continues to erode.  Growth is still unmanageable.   And IT stress levels increase (commensurate with CFO frustration levels) as the discussion remains on looming budget cuts that prevent buying the next new technology.

Often the answer is just behind the curtain.  In virtually every IT environment, there are process, policy and organizational problems that are precluding more efficient and coordinated spending.  Often IT can improve service levels, enhance data protection, simplify administration and ensure regulatory compliance … while at the same time dramatically reduce costs, and frequently without having to invest in new technology.

The most overlooked element frequently is the impact that organizational dynamics have on IT effectiveness.  As just a few examples of common organizational dysfunction that can undermine the best IT budgets and plans:

  • Groups that make procurement decisions to protect their budget.
  • New technologies that benefit one function, but are inconsistent with other plans, or incompatible with other applications or platforms.
  • Purchase decisions that are driven by a concern with job preservation.
  • Massive over-provisioning as a result of isolated technology purchases and capacity planning that doesn’t take into account other group’s plans or requirements.
  • Groups that make decisions that may unwittingly cause other problems – for example a virtualization initiative that may adversely affect security controls and decrease service levels.
  • Silos of decisions that are made without a ‘big picture’ view of the IT and business strategy.

The real challenge for IT organizations is to look beyond the technology solution and address the more difficult policy, process, and organizational issues.

Posted in Archiving, Business Services, Cloud Computing, Data Center, Disaster Recovery, IT Services Management, Virtualization | Leave a comment

Enough about Cloud already!

Posted on March 27, 2012 by Anthony Dickinson

Ok, I get it; enough about Cloud already!  In the recent weeks I've been to three Cloud industry events, Cloud Expo 2012 (London), The Legal IT Business Show—where I facilitated a panel discussion—and the Enterprise Cloud Computing and Virtualisation (ECCV) 2012event.  Nowdon't get me wrong, they were all well organised and informative events, but even for someone whosejob it is to stay abreast of what is going on the Cloud marketplace, I've heard about enough hype for now.

The presentations at these events seem to fall into one of two categories, they’re either interesting case studies from organisations that have been on the journey, taken the risks, and are realising the benefits; or are tired pitches from vendors and service providers touting the inevitability of Cloud and how the features of their product or service make it the Cloud technology to be trusted.  One interesting area of overlap, however, is the assertion that there really is no reason for organisations not to be realising the benefit in some way, shape or form.  The form it takes, however, depends on the nature of your business.

One thing that is clear from the case studies though, is that if you're likely to be realising any benefit soon, you probably started your journey some time ago.  And, you most likely have a business with a high potential for spiky demand.  I saw two great examples of this from Honda and Channel 4 (TV), with both of these examples relating to the provision of content to the consumer.

Channel 4, for example, is a mature user of Amazon EC2 that can launch new applications with zero CAPEX.  When they know from their TV schedules that the viewing public will be directed by the presenter to their website at 20:05, they 'dial up' from 5 web servers to 35 at 20:00 ready to absorb the peak; then back down to 5 again 23 hours later after the interest has passed.  Imagine the CAPEX wasted provisioning for such a peak using traditional procurement models.  Interestingly, however, while Channel 4 has virtualised its internal systems, it has no plans to use the Cloud for these.

Honda on the other hand limited their financial risk of a failed product launch.  In anticipation of the need for large quantities of data storage capacity, as the result of a digital media marketing campaign, Honda integrated Amazon's S3 storage service into their website.  The campaign failed to gain traction in the way that had been predicted, but Honda's additional storage costs were limited to £0.05, rather than the large CAPEX sum they would have wasted had they procured the estimated capacity of enterprise storage themselves.

Both of these examples show the benefits are real, but that they need to be applied with an understanding of the nature of the business activity and the potential benefits on offer.  Yet I still feel that rates of adoption are low, particularly within the enterprise.  Surveys by Microsoft suggest security, compliance and compatibility are the main barriers; the question is whether they’re too costly to address or just the excuse of legacy culture not wanting to change the status quo? 

That being said, on more than one occasion at ECCV 2012, organisations were advised to check the expense claims of their Sys Admins and IT Managers, as they will likely discover they’re using more Cloud than they think they are.  For News International, as an example, when they conducted this very exercise, the final total ran into seven figures!  This kind of Cloud leakage demonstrates that businesses are keen to unlock the benefits and if IT is unable or unwilling to provide them, they will be side-stepped.

Posted in Biotech/Pharmaceutical, Business Services, Cloud Computing, Data Center, Education, Infrastructure Operations, Other, Security, United States of America, Virtualization | Leave a comment

Think of the Cloud as an IT Tool? Your Competitors Don’t!

Posted on March 15, 2012 by Anthony Dickinson

No matter what your favourite definition of Cloud, and there are a few going around, it is simply the commoditisation of a set of activities across the computing stack as they transition from an ‘as a product’ to an ‘as a service’ world.  This process of commoditisation is driven by the basic economic forces of user competition (demand) and supply competition.  If an innovation is useful, it spreads as more people seek its benefits.  As its popularity rises, other suppliers improve on its design in order to grow market share–it becomes more feature complete, better understood and eventually standardises.

Commoditisation impacts all business activities, not just those in IT, and activities move through common stages from innovation, through customer build, product, commodity to utility services.  Therefore, the collection of activities that together make up your business can be categorised in this way.  This is useful, as it can help you understand how Cloud can enable competitive advantage for your business; and at the same time help you navigate the many Cloud service and deployment models available to you.

In the simplest of terms, business activities will fall into two groups 1) those that are strategically important in relation to your competitive advantage, most likely those activities that are more innovative and that you are doing yourself, and 2) those that are nothing more than a cost of doing business, most likely those activities that have commoditised that could be provided to you from a marketplace of providers.  Critically, each of these two groups benefits from different Cloud service and deployment models.

If you’re developing applications in support of your strategically important business activities then platform as a service (PaaS) can provide the benefits of componentisation and accelerate your time to market, enabling you to stay ahead of your competition.  This effect can be easily understood if you consider that, today a new website can be created in hours using an on-line service.  Imagine how long it would take, if each time we wanted to create a website, we had to design the CPU and every other component of the system.

For those activities that are a cost of doing business, software as a service (SaaS) can provide efficiency benefits, price arbitrage and minimise management overhead associated with an activity.  An often cited example in this space is Salesforce.com; few companies compete on how they do customer relationship management and sales force automation, so it makes sense to subscribe to an outside service.  Other examples might be human resources, payroll and email systems.  Of course, you still need to be able to trust the service will be available; imagine how difficult it is today to do business without email.

While PaaS and SaaS offer the ultimate benefit, getting there in a single step will not be possible for the majority of organisations.  This is where infrastructure as a service (IaaS) can be used to get the journey underway.  Amazon EC2 is perhaps the most well know example, but there are many other providers to choose from.  Combining IaaS with emerging ‘DevOps’ practices (the communication, collaboration and integration—often automated—between software development and IT operations) can deliver many of the benefits of PaaS and be easier to reach.

The key take away from this is that the value of the Cloud to your business, and therefore by implication your appropriate Cloud strategy, will be specific to each part of your business and depend on the profile of its activities.  In effect this makes Cloud a business tool and not an IT tool.  You can’t just procure and provide Cloud technology and expect your business to effectively realise the benefits.  A clear understanding of your business activity, and its relationship to the way your business competes within its marketplace, is critical to succeed with Cloud.

  • Anthony Dickinson, Service Director – Cloud Computing, GlassHouse Technologies

 

Posted in Biotech/Pharmaceutical, Business Services, Cloud Computing, Data Center, Education, Energy/Utilities, Financial Services, Infrastructure Operations, Insurance, IT Services Management, Manufacturing, Retail & Distribution, Telecom & Hi-Tech, Transportation, United Kingdom, United States of America | Leave a comment

Enterprise vs. Commodity Cloud, Why Care?

Posted on March 12, 2012 by Anthony Dickinson

 

The Cloud marketplace can be confusing to companies looking to take advantage of its benefits, yet minimise the disruption to the process, controls and security they’ve already put in place.  Before Cloud, the IT infrastructure marketplace was well segmented and well understood—you knew where to go if you wanted to buy networks, storage or servers.  In response to Cloud hype, practically every vendor and service provider has cloud-washed their products and services.  It is no longer clear to customers quite what they should be doing, or where they should be going to get unbiased analysis and help sorting through the deluge of vendor hype.

One choice appears to be between commodity and enterprise Clouds.  As the dust settles, we’re starting to see a number of camps forming.  Two such camps are the new breed of Cloud enablers and providers; this camp is pushing the commodity cloud.  The other camp is made up of the existing vendors and service providers; this camp is pushing the enterprise cloud.  To add to the confusion many vendors are offering products in both camps.  So what’s the difference and how does it matter?

Cloud is simply the evolution of a certain set of IT activities—the case for this is well argued by Simon Wardley of the Leading Edge Forum.  Therefore, no matter your favourite definition of Cloud, broadly speaking the Cloud phenomenon is the evolution of a set of activities across the computing stack, as they transition from an ‘as a product’ to an ‘as a service’ world.  The nature of any commodity or utility service in its marketplace is that it is both ubiquitous and standardised, and this leads to price becoming its defining characteristic.

While clouds have tremendous promise, as with any new technological shift, there are issues that arise.  The reality is that clouds fail, and IT organizations are left to deal with it.  A commodity cloud aims to deliver on the price promise and is designed to new practices, while being built from commodity hardware and open source software.  However, the past eighteen months has seen many of the household commodity cloud providers suffer from outage.  Even as recently as last week, Microsoft’s Azure Cloud was down for much of the day on the 29thFebruary—purported to be ‘a leap year glitch’.

Critically, you need to transition your architecture if you want to utilise commodity Cloud and avoid or mitigate failures.  The key point to understand when deploying your applications on a commodity Cloud is that, while they will reduce your unit costs, they WILL fail! and so your applications need to be architected to cope with failure—tolerance to failure needs to be designed-in from the start.  The problem is, most enterprise applications aren’t.

This architectural cost of change often kills the financial case and the allure of commodity cloud, this in turn leads to the enterprise Cloud.  These are designed to today’s best practice and built from more resilient and scalable hardware and proprietary software.  Built in this way, they are better suited to hosting today’s business applications and providing the necessary service levels required by your business; but, they can’t and they won’t deliver on the commodity price promise that is the logical conclusion of the process of commoditisation.  And, they don’t help you get there either.

Enterprise Clouds are still clouds and differ greatly from traditional IT provision.  While more expensive than commodity Clouds and specifically designed to support today’s business critical applications, to be a Cloud they must still deliver the five key characteristics, as defined by NIST:  on-demand self-service; broad network access; resource pooling; rapid elasticity; measured service.  By doing so, they differ greatly from the lengthy time-to-provision service-deck fronted shared services of today, and begin to unlock the sort after cost and agility benefits.    

Your Cloud strategy must provide you with a path to get your organisation through the architectural transition, and onto commodity Cloud, as soon as your specific circumstances will allow.  So be under no illusion, the architectural transition should only be a question of when and how, not if!  The rate of growth in data and information technology continues to accelerate, and perhaps counter-intuitively, all this increased efficiency only fuels the growth—this is known as Jevons’ Paradox.  As time passes, your IT estate is getting bigger, and the cost of architectural transition for your organisation continues to rise.  Act now!

  — Anthony Dickinson, Service Director – Cloud Computing, GlassHouse Technologies

Posted in Biotech/Pharmaceutical, Business Services, Cloud Computing, Data Center, Education, Energy/Utilities, Financial Services, Infrastructure Operations, Insurance, IT Services Management, Manufacturing, Other, Retail & Distribution, Telecom & Hi-Tech, Transportation, United Kingdom, United States of America, Virtualization | Leave a comment

Different Clouds for Different Crowds

Posted on March 5, 2012 by Jim Damoulakis

Is a ‘Commoditized Cloud’ really for everyone?   For most, sooner or later it will be.  Is it for every application?  Of course not. 

The recent Cloud Connect conference in Santa Clara served to sharpen the contrast between traditional IT and the cloud and should be taken as a wakeup call for enterprise IT organizations.  To me, the key messages that came through loud and clear in several of the keynotes and many of the breakout sessions essentially focused on:

  • The emergence of the commodity cloud as the most economically viable approach to realizing cloud services
  • The close linkage between the DevOps philosophy (or mindset) to automated systems lifecycle management and the cloud
  • The cloud as primarily a development platform for new applications (as opposed to a migration target for existing apps)
  • Cloud is following the historic route of adoption largely outside of (or despite) IT management and control

Whether you agree or disagree with these ideas, there definitely seems to be a coalescing of the cloud into distinct camps, with the two most prominent being the “commodity-crowd” versus what I’ll call the “traditional enterprise vendor” camp. 

Some view this cloud partitioning as between vendor cloud ecosystems, specifically between companies aligned with the OpenStack initiative and those aligned with VMware’s vCloud.   The implication – which doesn’t quite hold up – is between “open” and “proprietary”. 

For example, Amazon developed their cloud environment with their own proprietary APIs, largely because nothing else existed at the time. So they aren’t especially open, but they are representative of the commodity cloud.  There are also companies like Citrix and Cisco that would certainly be viewed as traditional enterprise IT vendors, but they also happen to be serious players in the commodity cloud.  Cisco is, of course, the “C” in the VCE (VMware, Cisco, EMC) alliance, but they are also a member of OpenStack and are helping to advance open networking initiatives like OpenFlow.  Likewise, the VMware-sponsored Cloud Foundry (http://cloudfoundry.org) is an open-source environment that has been well-received as a non-proprietary development platform.

Regarding the economic viability of one approach over another, it’s important to understand the reasons for considering a cloud approach, the specific benefits that the organization hopes to attain, and what type of legacy applications are in place. Some benefits, like improved utilization through shared resources, can be gained regardless of what kind of infrastructure and virtualization platform a company is on.  By developing standard offerings and being able to provision them in an automated fashion, an organization can derive significant benefit to users and improve IT efficiency regardless of the platform.  Architectures that provide flexibility can be crafted in a variety of ways and the efficiency of each approach is dependent on the capabilities and limitations of the given organization – the real key is intelligently leveraging your strengths and getting help for your weaknesses.

However, when considering scalability, I think it is no surprise that the “poster children” of the cloud – very large scale environments such as Amazon, Google, Netflix and others with massive server and storage requirements, developed commodity-based solutions to address their unique needs.  At such a level, the commodity cloud becomes a competitive requirement.   At Cloud Connect, Zynga described the growth of Farmville at launch as going from 0 to 10 million active users in 6 weeks!  This would have been nearly impossible in a traditional, non-cloud environment.

Clearly not all organizations are that dynamic.  If you have a small number of servers, hardware may not be the most significant cost driver or challenge.  Skill limitations may be the major pain point.  Even in larger organizations with the leverage to negotiate healthy vendor discounts, hardware may not be the driving cost factor.  Things like data center capacity limits, a preponderance of legacy applications, lack of flexibility and slow responsiveness may be the overriding concerns.  Implementing the appropriate cloud model in each case can help to address these issues, but the type of cloud will be quite different.  The former is likely to gravitate toward SaaS offerings that represent complete packaged solutions while the latter, depending on internal skills and purchasing practices, may seek out the solutions from organizations that can provide the vendor support that they depend upon. 

One major promise of the cloud is as an open development platform that can exist anywhere – internal or external, public or private – and is essentially indifferent to the type of platform it runs on.  That is great for new applications being built from scratch. 

To a large extent, the cloud direction for an organization will depend on its application strategy, the extent to which valuable legacy applications need to be supported, and the requirements of those applications. 

The commoditized cloud is a tremendous facilitator for the development of new applications, particularly those designed for massive scalability, and those that can provide resiliency within the application architecture.  However, applications that depend on highly available infrastructure do not necessarily benefit from the commodity cloud. 

While we are now hearing advocates of both camps expounding on the virtues of each, the reality is that no one approach is right for every environment or application, and we’ll continue to see lots of cloud types and variations for the foreseeable future.

- James Damoulakis is CTO of GlassHouse Technologies  

Posted in Cloud Computing, Data Center, Education, Infrastructure Operations, IT Services Management | Leave a comment

Live From Citrix Synergy with Atlantis Computing

Posted on June 1, 2011 by GlassHouse

Posted in Energy/Utilities, Financial Services, Insurance, Manufacturing, Retail & Distribution, Telecom & Hi-Tech, Transportation, United Kingdom, United States of America, Virtualization | Leave a comment

Live From Citrix Synergy with Liquidware Labs

Posted on June 1, 2011 by GlassHouse

Posted in Energy/Utilities, Financial Services, Insurance, Manufacturing, Retail & Distribution, Telecom & Hi-Tech, Transportation, United Kingdom, United States of America, Virtualization | Leave a comment

Live From Citrix Synergy with Log*in Consultants

Posted on June 1, 2011 by GlassHouse

Posted in Energy/Utilities, Financial Services, Insurance, Manufacturing, Retail & Distribution, Telecom & Hi-Tech, Transportation, United Kingdom, United States of America, Virtualization | Leave a comment

Intro to Citrix Synergy

Posted on May 26, 2011 by GlassHouse

Posted in Energy/Utilities, Financial Services, Insurance, Manufacturing, Retail & Distribution, Telecom & Hi-Tech, Transportation, United Kingdom, United States of America, Virtualization | Leave a comment