Table of Contents
- Summary
- Introduction
- What’s Available?
- Cloud Application Platforms
- Hypervisor-Centric Cloud Platforms
- About GigaOm
- Copyright
1. Summary
Although there is debate concerning whether internal (or private) clouds can actually exist, vendors and consumers alike have gladly taken to labeling their products and projects as “internal clouds.” They do so because many see cloud computing as encompassing a set of characteristics beyond the mere delivery of virtual machines over the Internet — it also includes features like on-demand self-service, broad network access, resource pooling, rapid elasticity and measured service. If most of these elements are present to some degree, it is both easy and accurate to catalyze discussions by invoking the cloud.
The problem this creates for would-be internal-cloud users is determining what capabilities they actually need from their cloud solutions, and what capabilities each internal-cloud vendor actually offers.
A good starting point for customers is determining the broad category of internal-cloud solutions that will best fulfill their needs, followed by examinations of the various vendors and products within that category. Because transitioning to a cloud-enabled environment can involve large degrees of technical, cultural and budgetary evolution, it is of the utmost importance that organizations deploy the right solution.
This report breaks the internal-cloud market into five distinct categories:
- Cloud application platforms: Rather than provisioning bare virtual machines a la Amazon EC2, cloud application platforms generally abstract altogether the computing infrastructure from the application and add resources or new instances as demand requires. Vendors in this space include Appistry, Elastra, GigaSpaces, IBM, Longjump, Oracle and TIBCO Software.
- Hypervisor-based clouds: These internal clouds consist of standard server-virtualization software fortified with management tools that enable cloud-like capabilities, such as load balancing, scalability, high availability, agility and automated provisioning. Vendors in this space include Citrix, Microsoft, Oracle, Red Hat and VMware.
- Internal infrastructure-as-a-service (IaaS) clouds: Internal IaaS clouds focus on infrastructure instead of applications. They let organizations’ IT departments act as service providers to organizational constituents — users self-provision virtual servers, run the workloads’ desire and IT charges them for the resources consumed. Vendors in this space include Abiquo, Cloud.com, Enomaly, Eucalyptus Systems, Joyent, MorphLabs, Virtustream and Voxel.
- Systems-management clouds: Large companies already go to them for everything from servers to provisioning software to governance, and now the Big Four systems-management vendors are attempting to cloud-enable their expansive software portfolios. Vendors in this space include BMC Software, CA Technologies, HP, IBM and Red Hat.
- High-performance computing (HPC) clouds: HPC clouds are ideal for those data centers in which large grids and clusters are already running compute-intensive applications. HPC cloud software acts as the intelligence engine, directing existing middleware products when and where to provision additional resources. Vendors in this space include Adaptive Computing, Librato, Platform Computing, Red Hat and Univa UD.
However, internal cloud software doesn’t operate in a vacuum, so organizations must consider other factors before deploying their clouds. Of these, hybrid-cloud capabilities — the ability to interact with public-cloud resources — and the underlying server architecture are among the most important. At this stage in the market’s evolution, most internal-cloud products offer some hybrid capabilities, ranging from mere application portability to the ability to manage externally hosted servers as part of an organization’s base resource pool. With regard to server architecture, organizations buying new cloud hardware need to consider whether commodity servers, blades or converged infrastructure serve as the foundation for their clouds. In some cases, server selection could affect software choices too.
In the end, regardless of how transformative internal clouds may be to IT operations, software is software. Organizations interested in deploying their own internal clouds will need to consider many of the same issues they do with traditional software purchases, including cost, lock-in, interoperability with existing solutions, compatibility with security/compliance protocols and even whether a vendor is likely to be in business several years down the road.