Understanding the Cloud Landscape

By George Reese
April 29, 2010 | Comments: 11

Making sense out of all of the components of cloud computing confuses even many of the major analysts. It's easy to understand how Google, Amazon, or SalesForce.com fit into the picture. But who is Eucalyptus and what do they do? Does CohesiveFT compete with enStratus or does it complement enStratus? And what is this vCloud thing anyway?

I've identified seven major components to cloud computing that I call "the seven pillars of cloud computing". I call them pillars because, as cloud computing evolves, they will form the foundation of a complete vendor cloud computing strategy. I'm not suggesting that you aren't a real cloud vendor if you lack one of the pillars—in fact, no one currently has a solid strategy for all seven pillars. The companies that end up seeing the most success as vendors in cloud computing, however, will not only develop strategies for all seven pillars, but also a strategy for integrating them into a comprehensive cloud offering.

George Reese is the author of Cloud Application Architectures: Building Applications and Infrastructure in the Cloud.

If you're involved in planning IT infrastructure as a network or system architect, system administrator, or developer, this book will help you adapt your skills to work with these highly scalable, highly redundant infrastructure services. Cloud Application Architectures will help you determine whether and how to put your applications into these virtualized services, with critical guidance on issues of cost, availability, performance, scaling, privacy, and security.

The Seven Pillars
Seven Pillars of Cloud Computing.png
The seven pillars of cloud computing are:

  • Virtualization
  • Storage
  • Cloud Orchestration
  • Data Center Systems Management
  • Cloud Infrastructure Management
  • Hybrid Cloud Integration
  • Public Cloud Services

As I mentioned above, there's no company that has offerings in all seven areas. If you are an enterprise looking for the complete cloud solution today, you may have to talk to as many as seven different vendors—though most likely fewer—before you find your answer.


Virtualization is absolutely not a requirement for cloud computing. Google and many others offer cloud services without using virtualization to make it happen.
From the complete cloud solution perspective, however, virtualization absolutely must be a strategic tool in the cloud vendor's arsenal.

Virtualization makes it possible to dedicated the minimum number of resources to a workload as is possible so you can more efficiently share real processing power among different workloads. It thus forms the backbone of many services we think of as cloud services today.

The number of players in this space is quite small. The significant ones are Citrix, KVM (Open Source), Microsoft, Oracle, and VMware.


From the cloud perspective, storage isn't about making the media on which data is stored. Instead, it's about providing the hardware and software for turning that media into a massive, abstracted storage tier available over a network and impervious to failure.

The current key players in this space are Compellent, EMC, Flexiant, HP, IBM, NetApp, Oracle, ParaScale, and a number of other smaller players.

Cloud Orchestration

There's actually nothing "cloudy" about virtualization. It simply facilitates the programmatic provisioning and de-provisioning of resources necessary for a cloud infrastructure. To make your virtualized infrastructure cloudy, you need an orchestration layer on top of it.

Orchestration layers hide the nuances of the underlying virtualization tier and exposes them to on-demand provisioning through web services APIs.

Companies with a footprint in this space include Amazon, CA (via their 3tera acquisition), Citrix, Enomaly, Eucalyptus, Flexiant, Rackspace, VMOps, and VMware.

Data Center Systems Management

Data center systems management is something most people forget about or ignore today when examining cloud services. It's the main tool a traditional IT operations staff uses to monitor provisioned resources, manage the provisioning, de-provisioning, and other IT workflows, and handle emergency situations.

The traditional data center management offerings are painfully "uncloudy" and the extent of the integration between traditional Data Center Systems Management and the cloud is largely limited to:

  • BMC monitoring of RightScale
  • enStratus pushing cloud health into Microsoft Systems Center
  • IBM's limited support for AWS in Tivoli

Everything else in this space is currently about supporting traditional data center management, provisioning, and monitoring activities.

This space is a well established space with some heavy hitting incumbents: CA, BMC, IBM, EMC, HP, IBM, and Microsoft.

Cloud Infrastructure Management

Cloud infrastructure management is really a specialized version of data center systems management. Cloud infrastructure management tools enable an IT staff to manage, monitor, provision, and de-provision resources in a cloud environment. In a few years, it likely won't be a pillar distinct from data center systems management. I break it out separately because there is almost no overlap between the two pillars today.

Cloud infrastructure management is dominated by start-ups today. The traditional data center systems management tools are poorly suited to the task of managing highly elastic cloud computing infrastructures. This gaping hole has enabled startups to come in and fill the need. On the other hand, there's very little overlap among the different startups in this space.

The cloud infrastructure management space is a very crowded space with a few major players: CA (via Nimsoft), CloudKick, Elastra, enStratus, and RightScale. In addition, CloudSwitch and CohesiveFT have elements of cloud infrastructure management, but generally fit more into the hybrid cloud integration pillar.

Hybrid Cloud Integration

Another pillar dominated by startups is hybrid cloud integration. Hybrid cloud integration services enable a company to glue together the different kinds of clouds supporting their infrastructure (public and private, multiple public, combined SaaS and Paas) into a single, coherent infrastructure. With hybrid cloud tools, you can manage and automate the movement of workloads among clouds as well as communications between components in different clouds.

This space is the least mature space with a very small set of players: CloudKick, CloudSwitch, CohesiveFT, and enStratus. Among these four players, there's almost no overlap in functionality. CloudKick provides a unified console for managing resources in multiple clouds. CloudSwitch enables you to securely lift your private data center/private cloud enterprise applications and drop them into a public cloud without changing IP addresses. CohesiveFT provides technologies for enabling virtual VPNs among clouds and managing virtual private clouds. enStratus provides governance, provisioning, auto-scaling, and auto-recovery across clouds.

Public Cloud Services

The public cloud services are what most people have in their head when they think "cloud". These are the SaaS, PaaS, and IaaS vendors that leverage the other components of cloud computing to deliver a public cloud offering. As a customer, you procure public cloud services on-demand and stop using them when you no longer need them.

There are so many players in the public cloud with so many different offerings: Amazon, Flexiant, GoGrid, Google, Microsoft, Rackspace, ReliaCloud, SalesForce.com, Terremark, VMware (via SpringSource), and so many others.

The Strategic Picture

The first thing that should jump out of this picture is how absolutely fractured the cloud computing space is today. No one has a foothold in more than four of the seven areas (unless you count VMware + EMC as one company, which I think is a mistake). All of the players with a traditional foothold in enterprise IT lack any capabilities when it comes to cross-cloud management and integration.

Because of this fragmentation, the cloud is essentially anyone's ballgame. A couple of players, however, have pieces that give them an early advantage:

Microsoft has developed strategies in three critical areas: virtualization with Hyper-V, data center systems management with System Center, and public cloud computing with Azure. In addition, they provide one of the main operating systems (Windows Server) that are used in virtualized environments in other people's public and public clouds. By fighting their traditional monoculture and adopting a hybrid cloud view, they could leverage these components to re-emerge as an innovative force in the computing world.

Like Microsoft, VMware has many of the basic pieces: virtualization, cloud orchestration via vCloud, and public cloud (both through SpringSource and partnerships like Terremark). They also benefit from the fact that many others are building their cloudy solutions around VMware technologies (Cisco, Terremark, BMC).

To this point, I've been discussing vendors in alphabetic order. I'm making a departure here because CA is emerging as a bit of a surprise. Generally known in the IT world as the place where technologies go to die, CA has managed a significant shift in just the past year as it relates to cloud. They've brought in a very pro-cloud management team, used their IT infrastructure management systems as a base, and begun making strategic cloud acquisitions in the form of 3tera and Nimsoft. CA is making all the right moves and the company is showing every sign that it "gets the cloud" strategically. All that remains is seeing how CA executes. It's possible in 10 years that we will recognize CA as the most important player in the cloud space.

A number of the "big names" are similarly showing few public signs that they "get it". Larry Ellison at Oracle rants against the entire concept of cloud computing while his sales force tries to convince everyone that they are a cloud player to be reckoned with (note to Larry: everyone's heard your rant and no one believes your sales force). Like Microsoft and VMware, IBM has all of the pieces of the puzzle but has made little public showing that they understand how to put those pieces together. Finally, EMC gets storage, but their strategic vision on its application to cloud is thin and wispy.

Other Cloud Stuff

There are other important roles with critical vendors playing those roles. As with any other kind of technology, systems integration vendors will be important in helping companies move into a cloud world. The companies leading the way today appear to be companies like Accenture, CSC, LTech, and SAIC.

Then there are companies like Cisco and F5 that simply don't fit neatly in the boxes I am trying to draw. These are hardware vendors that understand how their traditional offerings can be leveraged in a cloud world and have formed the relationships that will provide leadership in the cloud space.

Disclosure: I am the CTO of enStratus described above.

You might also be interested in:


Hey George - great discussion. Would suggest a couple of additions.
1) CA acquired Cassatt even before 3Tera. Being former Cassatt, I think they picked up a ton of good IC for a very small fee.

2) In terms of VMware and EMC, the V-C-N offering seems more mature and less costly.

3) You mentioned VPN-Cubed, but not VPC. Any thoughts there?

1. Really good point of Cassatt. That was really the starting point of "the cloud CA" and they got a lot of good people as part of that acquisition.

2. You're definitely right about the VCN (vmware/cisco/netapp) integration. It's one of the reasons I don't really think of EMC + VMware. VMware gets cloud, has a solid cloud strategy, and is executing on it. EMC is confused.

3. The value of VPN Cubed is that it is truly cross-cloud, whereas VPC is very much a proprietary solution. Still, VPC should be considered as part of a hybrid cloud solution.

Hello George,

Great Stuff.
I liked the way you classified the landscape to seven different segments.

As you rightly said, "there's no company that has offerings in all seven areas" and increases the complexity in adopting Cloud Computing for Enterprises.

One thing I felt missing in the whole stuff is Grid Computing and I am sure it can be an another pillar which solves many issues related to scalability. Though I am not part of Cloudera, I felt it should have been there in this landscape.

Good analysis, George. Like anything in this space, it's a current snapshot given who's doing what at the moment.

I think it's difficult to find a way to talk about all the pieces that are part of thinking about cloud computing and have some apples-to-apples way to categorize them. As you noted, the data center systems management and cloud infrastructure management are similar in concept, but vendors are in very different stages of delivery on this (our Nimsoft acquisition has certainly been a big boost for CA as you noted).

Also, I was going to quibble with calling out virtualization as a pillar, but then your first comment on it was something I agree with: "virtualization is absolutely not a requirement for cloud computing." It's an important technology and can certainly be a big help in cloud computing, but it's not the strategic issue. So, I'll let it slide. :)

@jayfry3 (also a former Cassatt guy)

I read it as a great analysis, and what you wrote makes sense.

I also try to see things from the buyer side (the one who pays money to all these cloud vendors, so they should be important, I guess), and I read before from a separate analysis, there are Content Consumers and Content Contributors.

Consumers are literally ones who use the cloud as information source.
Contributors are ones who create contents. Data, applications, etc.. whatever it is.

The simple analogy is that contributors are interested in IaaS, PaaS, and consumers are interested in SaaS.

To generate a set of pillars that attract the most contributors, will organically become a great attraction for consumers.

As simple as it sounds, I also would say it's not always easy.


That's really the strategic question that I'm opaquely getting at. How do you organize what you do in the cloud to be able to support the largest range of customer needs?

It's looking more crowded and specialized, George that it seem difficult to have more cloud providers, especially small names like Morphlabs, to enter the market and get a piece of the pie.

But then again, there'll be plenty of customers still waiting in the sidelines.


I think the Cloud is a integration of quite a lot of management,system,and detailed applications.So it's really a special market.
I am also planning to build a private cloud to provide several key services.We don't want to carry out the whole program only after we make a complete plan because nobody will know what is a complete plan.hehehe.
Oh yes,the only agreement that we are sure is that we will work under vSphere.

Learn more and get started on the Cloud for free

When you join the Front Runner program, you can access one-on-one technical support by phone or e-mail from our developer experts, who can help get your applications in the cloud. Once your application is compatible, you'll get a range of marketing benefits to help you let your customers know that you're a Front Runner.

May I suggest you check out Surgient? You mentioned several startups in the space, but only Surgient has offered a product in this space since before the term "cloud computing" became popular (product introduced in 2003, several "cloud" patents granted).

I work for Surgient, so I won't expound on the solution's advantages here, but I hope you give us a look for future articles.

Mike A.

Can you comment on Jason Scott's essay concerning the future of the cloud?


Given the notable failures that have already occurred in cloud offerings, what contractual obligations do you believe individuals and companies should extract from cloud vendors prior to depending on them for any services whatsoever?

News Topics

Recommended for You

Got a Question?