Mega-awesome BlueLock engineer Bill Barbour recently posted a great snippet from marketing guru Seth Godin on the BlueLock quote board:

This Seth-ism is true not only with user interface design and information overload, but also reflects how the constant bombardment of tools that promise to make you "cloud-ready" can make this whole cloud technology thing seem needlessly complex. The current media coverage and blog blitzes about creating cloud-ready applications can make you feel woefully unprepared if you don't have a build grid dedicated to continuous integration, on-demand provisioning via some programmatic first-order language, a MapReduce cluster, a 100-node NoSQL database fabric and an immense automated dashboard notifying you to every quantum tunneling event.

It doesn't have to be this hard.

I think a good deal of software engineers and infrastructure gurus are excited about the possibilites that accessible cloud technology brings. That excitement has powered a ton of innovations, notably the current devops revolution that is accelerating software engineering and delivering applications faster than ever before. We now have applications like Pallet and Chef that can provision and re-provision cloud computing infrastructure in a dependable and repeatable manner. However, just because Pallet and Chef now exist does not mean you have to use them in order to fully leverage a cloud hosting strategy.

If you can move faster with bash scripts via SSH, go with it. If your provisioning strategy consists of deploying the same vApp template over and over again, rock on. If you can do configuration management just as easily with Git or Subversion, have at it. If you wish to provision your cloud computing infrastructure with the same slipstream ISO used for your physical hardware, do so to your heart's content. One of the advantages of cloud computing is that you have the latitude to manage your hardware as you see fit, unbound by the constraints of a capital expense form every time you want to expand. Create your own Linux distro, load up enterprise applications from the 1980's, build a virtual machine from a series of floppy disks.

One of the advantages of BlueLock's VMware Cloud is that you can begin hosting your applications just as you always have, even if you're accustomed to a physical data center. As your needs expand and you wish to try out different devops approaches, you can easily do so with little risk. With cloud hosting you can manage your infrastructure however you want while still having room to try something new and accelerate your applications when you feel the whim.

Near as I can tell, Azure was launched under the premise of a .NET Platform as a Service offering. Microsoft employed some clever tricks to provide a cloud-ready relational database and provisioning tools for Visual Studio, bringing the service nose-to-nose with Google App Engine and Salesforce.com.

Bear in mind that Platform as a Service is quite different from Infrastructure as a Service. PaaS offerings focus almost entirely on the developer, leaving scalability concerns entirely to the service provider. IaaS provides a more familiar server-based architecture, allowing IT to manage infrastructure concerns as they see fit. With PaaS the secret sauce lies with the "automagic" scalability, almost akin to grid computing in some facets. PaaS providers seldom (if ever) dabble into IaaS offerings, often because the data center layouts and architecture are vastly different.

Imagine the surprise when the Azure team announced their "Virtual Machine Role," a Windows Server 2008 R2 virtual machine running in the Azure cloud. This casts Azure as a weird hybrid PaaS/IaaS platform... and given how disparate those services are I'm not confident they can combine well. This either means a) Azure has always just been a league of Windows VMs or b) the ability to have individual virtual machines has somehow been wedged sideways into the platform. Either way it doesn't instill a whole lot of confidence in the PaaS solution. I can see maturing from IaaS into PaaS, but the other way around seems odd to me. Maybe Azure started as IaaS and then became PaaS and then became IaaS?

My head hurts.

I concede that I may just be a jaded purist. Still, I can't help but think that Publilius Syrus had the right idea about cloud computing services: "to do two things at once is to do neither." I wouldn't assume to say a service provider can't provide a choice between the two, but it seems awfully strange to wedge a virtual server into a PaaS offering.

Recently I had the pleasure of representing BlueLock at a reception for Mira Award finalists. Yes, this cloud computing company is up for the prestigious Gazelle Award. At the event, each finalist company was asked to provide a 60-second elevator pitch on what it does. Since our own branding work is still in progress (I hired a branding firm to help us), I gave it my best shot.

I began by first describing BlueLock as a cloud hosting company and referenced that we have spent our lifetime in the cloud, even before anyone knew what cloud computing was. Deciding to engage the audience, I asked a lawyer friend of mine to share what the cloud is. His response was, "Kim, everyone knows what the cloud is" and he pointed out the window.

Since I haven't yet come up with a simple definition of cloud computing that my own mother will understand, I have found it helpful to use an analogy. (Thank you, Nicholas Carr who wrote The Big Switch!) I begin by sharing that years ago companies had to generate their own electrical power and spent considerable manpower and other resources doing that. Their ability to generate that power served as a competitive advantage at least for a window of time. Later on companies were then able to simply plug in to the electric grid, access as much electricity as they needed at any given time, and paid for only what they use.

The same is true for IT today and that is what the cloud is about. At BlueLock, we help companies focus on and grow their businesses by providing them with the amount of IT infrastructure they need, when they need it, and charge them for what they use. And this can all be done in minutes via the Web. This serves as a tremendous value to our clients who would prefer to spend their time and resources on growth and revenue generating activities, not servers and other technology infrastructure.

I am still on a mission to come up with that simple and clear definition of the cloud.  Any and all input is welcome! For now, the electricity analogy seems to work. My mother gets it. What will also work for me is winning the Mira Award on May 15.

Stay plugged in for that!

This news warms my heart.  As an Indiana University Bloomington grad and now working for a leading Infrastructure as a Service provider (cloud provider) I'm ecstatic to hear that my alma mater is heading to the clouds in a big way.  Let's not be foolish to think that IU has been out of the cloud party completely, the IU School of Informatics is on top of their game.  But now, with the news of a grant specifically dedicated to cloud computing, IU should be making some big leaps in the arena. 

Indiana University has received a $1.5 million grant from the National Institutes of Health to develop a software infrastructure and use cloud computing for a variety of life sciences and genomics data applications, IU said today.

"This research is potentially path-breaking," Peter Cherbas, a professor of biology and director of the IU Center for Genomics and Bioinformatics, said in a statement.

"Cloud computing approaches are likely to change the nature of our national research computing infrastructure in the coming years," said Principal Investigator Geoffrey Fox, director of the Digital Science Center and associate dean of research and graduate studies in the IU School of Informatics and Computing. "These technologies hold significant promise in the life sciences and medical sciences as they offer the potential for greater computational power and faster speeds at a lower cost, and in a way that is easier for scientists to use than traditional grid computing approaches."

"Contemporary DNA sequencing machines are churning out data at rates that would have been unimaginable to biologists just a few years ago,” he added. “To use these data — to turn data into some kind of understanding — will demand good tools for using the Cloud and those tools will impact genomics projects worldwide.”

Ahhh...the joys and economics of cloud computing.  This is what its all about - churning lots of data at a fraction of the cost and in a fraction of the time.  Click here to learn more about how we do just this for our clients.

Click here to read the news release about IU's Cloud Computing grant.

1. It’s not all-or-nothing move, most will opt for a hybrid approach.
Cloud computing will not necessarily replace a company’s entire infrastructure or serve as a total outsourcing option for data center operations on-site.  Most companies will use the technology in addition to their resources, blending public and private clouds to maximize efficiencies and resources.  I would even argue that companies will begin to adopt different cloud providers for different business functions (i.e. test/dev vs. production environments) because of differing needs and requirements in addition to their on-site resources.

2. Workloads
Some argue that cloud computing is best used for pre-production workloads such as test/dev and storage rather than production.  I’m here to say that some providers are able to handle different workloads better than others.  For example, BlueLock offers 99.99% uptime in the cloud because our clients are enterprise-level companies with mission-critical production environment needs.  They can’t afford the downtime that would be threatened with a provider who can only promise 99.9% or 99.95% uptime.  But similarly, clients who only need test and development models don’t need to pay the premium for a 99.99% uptime cloud environment when they only require 99.9%.

3. One size doesn’t fit all in the cloud
Each company has its own needs and requirements when it comes to their computing environment.  A financial company cannot and should not use the same type of an environment a social media site does.  A cloud computing model or environment must be built to service the specific needs of that client.

4. It’s just a delivery model
Cloud computing isn’t a technology – it’s just a shift in the way the world does IT.  Cloud computing represents a combination of technologies such as SaaS, online storage, and grid or utility computing.

5. It’s for the big guys and the small
Many think cloud computing is just for the SMBs, but many large corporations have jumped on the bandwagon as well.  BlueLock has added on a number of large enterprise clients this year who are using the cloud because it makes the most business sense.  Outside of BlueLock, the market is seeing large companies consume a mix of cloud services – public, private and both.

6. Self-service
Self-service is one of the cornerstones of cloud infrastructure.  Some transactions just don’t require “people” to be there and the logical step was to remove that element in the service.  Therefore, the self-service aspect of cloud computing is a huge step in automation and cuts costs on both the buyer and seller sides of the hosting business.

7. Standardization
Cloud computing is highly touted for its ability to create efficiencies by cutting out the capital expenses and transferring those to lower operating expenses (subscription or pay-per-use model) and for its on-demand benefits.  But that’s not the big picture – cloud computing allows people all over the world to access, connect and work smarter, faster, better. 

Cloud computing providers differ in two main ways: people and control.  No two cloud services are alike for that very reason and people and control are the two variables that differentiate cloud computing providers from one another.

The definition of cloud computing varies everywhere, but some main characteristics of “the cloud” are: virtualization, capacity on demand, shared resources and subscription-based pricing.  If you fit that bill you’re a cloud, but which segmentation of cloud?

It's either:

SaaS (Software as a Service) – The web-facing software that users interact with (i.e. SalesForce or Google Apps)
PaaS (Platform as a Service) – a platform that helps utilize cloud resources (i.e. RightScale or rPath)
Or IaaS (Infrastructure as a Service) – the actual infrastructure behind the cloud, the virtualized servers and network (i.e. BlueLock, GoGrid or your internal cloud)

For purposes of this post, we’re going to focus on Infrastructure as a Service (makes sense because that’s what we’re experts at).  From here, there are also a few different type of IaaS clouds (surprise, surprise!) and their characteristics, to make things just a little more interesting.  Here they are:

Private internal cloud:
-    Limited Capacity
-    Expensive
-    Highly Secure
-    Enterprise Features
-    Self Managed

Public utility cloud
-    Limitless Capacity
-    Lower cost
-    Self Managed (DIY)
-    No/DIY SLA
-    No/DIY Security

Full-service public cloud
-    Large Capacity
-    Fully Managed
-    Customized Security
-    Enterprise Features
-    Enterprise Services
-    SLA

Back to IaaS cloud differentiation – again, all about people and control.  First, you have to decide how much control of the infrastructure you want, can manage or more importantly, can afford.  Then take a look at the other components that will affect your costs in the cloud (some of these you may already have):

-    Hardware & software
-    Licenses
-    People
-    Place

If you’re looking at a do-it-yourself platform like Amazon EC2, you’ll have to purchase your own people and they’ve have to monitor, control and manage the infrastructure.  With this option you’ve got a lot more control, but a much higher TCO – because those people are going to cost you lots of money.  But if you don’t have the people, or you want them to focus on something else, like your core business (i.e. your software solution) instead of the infrastructure then a trusted full-service cloud option like BlueLock makes a lot more sense and could save you about $50,000 a year. 

How’s that even possible?  When you purchase a BlueLock solution, you’re not only purchasing the infrastructure and place to house that infrastructure, you’re purchasing the people and expertise behind it – meaning less employee costs for you (or more beneficial employee costs).  Some would argue you lose a little control because your people, your team isn’t the one working on the infrastructure, but others think that a partner like BlueLock is just an extension of their own team – and a better use of their IT budget at that.  Take Projetech for instance – watch this video and you’ll see that they really see BlueLock as an extension of their company not just another vendor, and that’s the way we like to think about it too. 

I would argue that it’s not just the future of the internet; it’s the future of computing as a whole.  Del Diax recently wrote an article entitled, “Cloud Computing, the Future of the Internet” in which she examines the pay-as-you-go computing utility that is saving companies thousands of dollars each month.

No longer do companies have to engage in a huge capital outlay for the servers, racks, and even people required for a traditional data center.  They are now able to rent computing power and pay only for the compute (and expertise in the case of BlueLock) they use. 

Diax interviewed Matt Mullenweg, founding developer of WordPress who claims, “The biggest mistake we made at WordPress.com in terms of infrastructure was buying servers.  Now we lease them all month to month.”

BlueLock started talking about this infrastructure as a service idea a few years ago (when it wasn’t nearly as popular as it is now) and the trend towards outsourcing and renting servers and the licenses to run infrastructure has really caught on, specifically with the economy troubles that we’ve encountered.  It just makes financial sense. 

Cloud computing can be grouped into 3 categories – Infrastructure as a Service (IaaS), Platform as a Service (PaaS) and Software as a Service (SaaS). 

IaaS provides the infrastructure – and at BlueLock that includes the management and monitoring (people) required to run the infrastructure.  Amazon’s EC2 and other players like GoGrid and RackSpace are providing just the infrastructure.  No matter what level of infrastructure service you’re looking for, I think it’s become quite evident that the cloud, in all its grandeur, is really proving itself as the future of the internet and computing as a whole.

Rob England recently wrote an article on cloud computing entitled, “Cloud Computing Outlook Far From Sunny.”  He adamantly argues that cloud computing is a good fit for “high risk/low-capital applications like startups or small business or websites” but has a “more jaundiced view” towards the cloud for enterprise computing and existing core applications.

Yes, everyone has jumped on the cloud bandwagon and is touting “cloud” on every offering possible in hopes of grabbing a small piece of a very large pie that being eaten up as we speak, but that doesn’t mean it’s a worthless technology for the enterprise.  England says, “Quite simply, the idea is impractical for legacy enterprise applications.”  He claims that this “technical solution” doesn’t really solve the non-technical business problems, but actually presents more problems, introducing greater complexities to manage.  He doesn’t have software as a service in mind in this article, which is good because neither do I.  He is talking about the “internal grids or hosted computing or the myriad of other things that seem to get lumped into ‘The Cloud.’”  Great, because I’m assuming by that, he’s including what we call “infrastructure as a service.” 

The number one problem, he argues, is migrating legacy applications.  It is true that some applications simply weren’t built for the cloud, which is probably why he is so skeptical of legacy migration.  But many others would argue that it won't be impossible forever.  Bernard Golden, author of The Case Against Cloud Computing, believes that at some point someone will develop a physical-to-cloud migration tool that can alleviate those technical migration pains.  I’m with him.  It’s going to happen. 

But more importantly, I don’t believe that there is any reason why enterprise clients should stay away from cloud computing.  There are many enterprise-level cloud options out there.  And if the worry is that there will be too many pains associated with a large enterprise moving into the cloud, why don’t they opt for a full-service cloud, like those offered by managed service companies (such as BlueLock)?  Then they don’t have to worry about the resources, people and time it takes to migrate everything to the cloud, their service provider will do that for them. 

Jake Sorofman, our friend at rPath, wrote an interesting piece in Dr. Dobbs.  He sings Amazon praises, and says, "today, EC2 may well be the truest expression of cloud."  Yes, according the numerous definitions of cloud, Amazon most always fits the description.  And Jake states, its "just a darn good example of what a cloud should be."

But in the cloud computing realm, there needs to be competition and innovation and choices among cloud computing providers that can drive the messages and solutions around the advantages of virtualization. 

Jake writes:

For cloud to marute into mainstream and potentially transformative alternative to traditional computing, there must be many players -- each contributing their own unique value to the marketplace.  Luckily, we've seen innovators like Rackspace, GoGrid, BlueLock and private cloud infrastructure like Eucalyptus deliver their own unique pieces to the cloud puzzle.  We'll also see major platform vendors like IBM, EMC, and Oracle/Sun with their own cloud offerings.

I thnk its becoming more and more accepted that virtualization and cloud computing will become more of the norm, and what will matter the most during that "switch" is the ability to control and deploy the systems that run the infrastructure.  What Jake doesn't mention is that rPath is innovating ways to do this every day.  Without the virtual appliances they are creating like the rBuilder and the competition that will come their way, he's right, the cost and risk of cloud "may actually wash away the benefits of this new computing model itself."

Click here to read the full article.

A recent post by EDL Consulting comments on what kind of an edge the first movers in the web-based hosted software solutions will have in regards to the cloud computing market.  Will they have an edge or will they be swept up by larger companies with large existing client bases like Microsoft?

With the advantages of virtualization becoming more and more widely known, companies are jumping on the cloud bandwagon quickly.  Reuters has suggested that even those cloud companies that got on the bus early will still face fierce competition from larger companies that will come in later with huge databases of trusting customers.

While the post is focused on the Software as a Service aspect of cloud computing, I think that the discussion rings true among cloud computing hosting providers as well.  BlueLock has been doing infrastructure-as-a-service for two years now, before the big cloud boom hit.  Suffice it to say, we definitely feel that we were one of the first movers and shakers when it comes to managed cloud hosting, especially considering the fact that when we first started talking about the value of virtualization our prospects looked at us like we were crazy.  Crazy or not, BlueLock was built around the value of virtualizing IT infrastructure to shift capital expenses to a monthly operating expense that's a fraction of the cost (plus we could throw in virtual disaster recovery for almost a 3rd of what it would cost in a traditional model).  So maybe we were crazy, but the idea has caught on.

In terms of advantages and disadvantages of being a beginning mover and shaker in the cloud computing arena, we'd have to say that BlueLock has definitely found some advantages.  We've seen our competition try to add on cloud services to their traditional hosting services, and in many cases its a difficult transition.  How do you know what to tell your salespeople?  Sell cloud or sell traditional colocation?  The companies had worked for years building a business on traditional hosting services and then virtualization comes along and shakes everything up - its less expensive, there's more space and compute to work with at a fraction of the cost, etc etc.  Your margins and MRC start to decrease, but you need to keep up with the cloud competition.  Not only that, but you don't really specialize in virtualization, the traditional way was your thing.

Well at BlueLock, we started with virtualization and that's what our engineers specialize in so its a no brainer for us what we should sell and its a no brainer for us on how to provide the biggest bang for our prospect's buck.  As for the bigger companies cramping our style - the Rackspaces and GoGrid's of the world aren't really our direct competition anyway. 

We surely feel we've got an edge.

April 7th, 2009 : Rich Miller

Over the past week, there have been a series of electronic attacks on major Internet web hosts and domain service providers. These distributed denials of service (DDoS) attacks have disrupted service for tens of thousands of web sites. Here’s a recap of the recent activity:

Coincidence? Conficker? Obviously, there’s no way to know. But a worrisome common thread is that these were all large-scale attacks that disrupted providers who’ve seen plenty of DDoS attacks before.

According to an article on Data Center Knowledge, many companies are turning to the cloud to reduce their capital requirements, others are evaluating co-location and leased space in data centers.

At our data centers in Indianapolis and SaltLake city, BlueLock can offer clients either alternative; traditional collocation or cloud computing solutions.

In 2006 AFCOM  produced a report based on extensive research including five very bold predictions for the future of data centers.   This week, Rich Miller took a closer look at the original predictions three years later. 

I was most intrigued by Prediction #4: By 2010, nearly 70 percent of all data centers will utilize some form of grid computing or other virtual processing.

Virtualization has been adopted even faster than expected by the DCI 2006 findings. AFCOM’s survey from November 2008 found that 86 percent of data centers expect to increase the use of virtualization. The enthusiasm for virtualization hasn’t yet translated into a shift towards cloud computing. More than 77 percent of AFCOM respondents said they were not planning to increase their use of cloud computing.

With one year left before the clock runs out on this prediction, what do you think?  Will the current economic crisis drive the change faster?

As companies move from enterprise colocation and traditional data centers to cloud computing environments, the roles of their information officers and managers will change.   In an article for Computer World, Mark Evert Hall describes the changing role of the CIO.  He says:

“The CIO's job is changing again and, I suspect, we'll start seeing new titles to go with it. The contemporary IT leader will be charged with overseeing an organization's computer-dependent services in addition to everything else. For lack of a better TLA, (Three Letter Acronym) I'm dubbing the new position with four letters, the CCSO: cloud computing services officer.”

What will these new CCSO’s do?   Primarily their role will be to oversee the use of subscription services, not just SaaS but access to on-demand IT utility grids as well.   With the technology evolving at such a rapid pace, this is no position for someone uncomfortable with change. 

Who will be your CCSO?  What plans do you have in place for training and grooming new managers for these dynamic roles?

Rich Miller raises an interesting question in a recent blog post: Can Cloud Computing handle compliance?  He does a good job of summarizing the conversation between Chuck Goolsbee at SearchDataCenter.com and Michael Sheehan from GoGrid.
  
In their conversation they touched on many of the common issues and concerns regarding, security, PCI compliance, infrastructure, software configuration and network deployment. Their discussion also included their opinions surrounding the cloud computing business model.
 
The article is filled with links to more conversations and discussions, all worth reading.  The bottom line: There is a general theme supporting the belief that virtualization can be compliant. At BlueLock we work closely with partners VMware, Shavlik and others to insure a secure, stable and compliant environment for our clients.  We believe this will be an ongoing conversation in our industry being that there are no set compliance "rules" around Cloud Computing. 
 

 
There are numerous definitions of cloud computing currently circulating.  Here is one from JDJ  I like because it compares Cloud Computing to something everyone understands; their television. The author says:

"Cloud computing is to storing and processing data what the electrical grid is to plugging in your television: a scalable way to deliver services while matching supply and demand across the grid."

To this I would add the following: The average consumer doesn't really care how or where the electricity is generated to power their television.  They don't care about utility company loads.  They simply want power when they turn on their TV. 

The same is true for IT managers and business owners.  It is less important where the data center is located, Indianapolis, Salt Lake City or somewhere else.  What is important is when excess capacity is required, the capacity is there.