There are hundreds of definitions of what the cloud is, but at its most basic, cloud computing is about the ability to enhance enterprise value through more efficient use of IT resources. As with any other IT solution, the first major step when considering a cloud-based solution is to understand the value it can bring to your enterprise or government agency.
Some organizations just want to move from the infrastructure they own to the infrastructure they rent. For others, whose operations are transaction and data heavy, with many integration points, transitioning is much more complex, and involves advanced concepts like dynamic server provisioning based on usage, moving virtual machines between multiple cloud providers, CDMI support, etc.
Elasticity is a fundamental factor when considering a cloud computing strategy, and a major thorn in the sides of CIOs, who grapple with servers that operate at as little as 3 to 4% of capacity. There is a critical need for enterprise to have capabilities to rapidly scale up or scale down to support changing needs.
In other words, companies want the ability to utilize cloud services in a fashion not unlike a utility, paying only for services and products that they need, exactly when they need them. Cloud providers thus earn their revenues based on their ability to “burst” on demand – providing organizations the opportunity to readily access virtual and scalable IT resources (both human and system.)
These scalable “bursting” scenarios (aka elasticity) range from virtual machine instances for a small start-up company to a sizable footprint across data centers in multiple locations to provide transparent failover capacity for a large company during a disaster.
Cloud Models: Which one do I need?
There are three types of high-level cloud computing models that provide various benefits to meet differing needs. While there are still not many well-defined standards when it comes to choosing a cloud solution, most often the decision will be driven by compliance needs, price and complexity.
This model is owned and managed by service providers on a metered, or services used, basis. Offerings are generally standardized yet flexible for rapid deployment and cost efficiencies.
Public Cloud Advantages:
- No capital outlay
- Fast deployment of standardized processes
- Limited commitment
- Metered/Services payment model
- Most elastic model
Public Cloud Disadvantages:
- Can be less flexible; generally fixed/standard services
- Security threats
- Varying availability and service level offerings across providers
- Questionable ownership of data
This model is generally (but not always) privately owned and managed by the organization and provides a highly secure (yet available) environment for internal and partner use. Private clouds can be implemented in an on-site, internally managed model or as a provider-hosted cloud with dedicated resources (which is why the term private cloud can seem so “cloudy”.)
Private Cloud Advantages:
- Potential for less capital outlay than internal data center
- Flexibility for customization
- Tighter security/privacy
- Readily auditable; data ownership
- High level of control over environment
Private Cloud Disadvantages:
- More costly than public clouds
- Capital expense brings higher commitment
- Less elastic model (yet more elastic than traditional internal data center)
This model fuses the public and private models to meet differing needs for ownership, management, and security/accessibility across business processes and applications in an appropriate mix.
Hybrid Cloud Advantages:
- Combination of Public and Private advantages based on optimal flexibility of delivery at most efficient cost structure
Hybrid Cloud Disadvantages:
- Overall complexity of environment
- Increased environment management/control requirements
Management must carefully determine the approach for moving any part of their business to the cloud based on short-term and long-term needs as well as their overall requirements for security, accessibility/availability, and cost. Even the largest of organizations with a need for tightened control, security, and auditability may find a gradual entry to be most optimal. This could mean:
- Utilizing public cloud Software as a Service (SaaS) features for small projects and applications;
- Moving Intranet and traditional data center infrastructures to a highly secure private cloud via Infrastructure as a Service (IaaS); and
- Steadily hybridizing the Cloud model to reach maximum benefits with minimal downside while still meeting all requirements for security and flexibility.
As always with technology, evolution is mandatory and management must consider the services and benefits offered by the cloud to remain competitive. The simplest starting point is to consider your business needs for scalability in a very dynamic, agile nature, and the need to create greater cost efficiencies.