Cloud A-Z Part 4: The complexities of IaaS exposed

AvatarWritten by | Uncategorized

Infrastructure as a Service is in a state of constant evolution, which makes it very challenging for most businesses to understand. They don’t understand how to leverage it or how it impacts a move to cloud computing. That’s why one of the most important services a partner can offer is IaaS. It allows your customers to move to the cloud without having to understand the underlying hardware and software necessary to support their core business in the cloud.

The Vision and Economics of IaaS

The idea behind IaaS is to be able to offer enterprise-class data centers to end-users regardless of their size and financial status. They will have access to flexible storage that can be provisioned at a moment’s notice to meet ever-changing needs and the support structure behind it that would normally be beyond them. It allows businesses of any size to change their infrastructure to meet new needs without expensive initial payouts, as well as eliminates the inflexibility of an on-premises solution.
As noted in Cloud A-Z, Part 2: Cloud Computing’s Economic Value, there are multiple economic benefits to moving away from on-premises and onto Infrastructure as a Service.

  • Lowering the opportunity cost of running technology
  • Lowering the total cost of ownership of technology
  • Shifting focus to core business activities
  • Moving IT costs from capex to opex


Capacity vs Utilization Curves (Source:

The Building Blocks of IaaS
Infrastructure as a Service is the most basic of the three types of “as a service” offerings typically thought of when considering the cloud stack. It usually consists of:

  • Compute
  • Storage
  • Network
  • Cloud databases
  • Monitoring and autoscaling

Like a computer, it’s probably easier to buy the complete offering, already built; but if you know what the individual pieces do, it’s much easier to build a custom solution that will meet your customer’s needs exactly.
Compute is the ability to leverage servers on-demand. Most of the characteristics you would consider here are not much different from your regular computer or server – operating system, RAM and disk. But there are two components that are different: the hypervisor and virtualization.
A hypervisor allows multiple operating systems to run simultaneously on a single server. This software allows for virtualization, which enables multiple pseudo-servers to run on a single physical device. These two tools, but particularly the hypervisor, ensure the servers are provisioned properly, and no single user corners the resources.
The operating system is usually determined by the applications that are going to be run upon it, as well as personal preference. RAM, as most know, is a type of data storage that’s used when a program or document is read from a hard drive. Again, this is usually determined by the type of applications that are going to be run. Typically, you will want to choose the smallest amount, or slice, that you can run the job on because you can always enlarge your slice instantly, as needed. The disk provides the storage available locally for such things as loading an application onto a server. Like the other two pieces, what is going to be run on the server should determine the size of the disk.
Storage consists of two pieces: the cloud storage and the content delivery network (CDN). The cloud storage can either be block-level, similar to attaching an external hard drive to a PC, or file-level storage which stores individual files and doesn’t require mounting a storage block directly to the virtual server. File-level storage is great for simpler storage needs like images, videos and scripts.
The CDN may not always be available from a cloud provider, but it’s valuable to have. It consists of various servers placed in different locations with copies of the same data. The goal is to optimize distance for the customers. The server that is closest to a customer will deliver the data.
Network components are another consideration. For a business to implement its own data center, it has to invest in the hardware and software, but also the networking needed to tie the servers together. There are several pieces to creating a network: switching, routing, domain name system (DNS), and load balancing.
Switching and routing determine, respectively, how data is grouped together and what path it takes to get to where it’s going. There are significant costs inherent in purchasing the most current technology available and keeping it up-to-date to ensure the speediest rates of transfer.
The DNS refers to the naming system that is used to identify everything connected to the internet. It’s the equivalent of a phone book and ties a name to each IP address. The DNS server, for instance, determines which IP you connect to if you type into the address bar. Cloud DNS services allow for managing all aspects of the user’s DNS records.
Load balancing is a technique for distributing traffic across multiple servers. It prevents clogs, delays, etc., in data transfers. A load balancer can increase capacity and reliability.
Cloud Databases
Cloud databases work like other traditional databases. They’re scalable and configurable creating an easy way for an organization to create, access and store their data. Again, database as a service (DBaaS) eliminates the need for an organization to invest in hardware and software, updates and personnel to maintain the application. When combined with the other elements of IaaS, you can create a very customized offering for your customer.
Monitoring and Autoscaling
Monitoring and autoscaling give visibility into system activity and can automatically provision the account to ensure your customer can meet their user’s demand. Monitoring is a big piece for your customers as many are afraid they will lose visibility into their infrastructure if they move to the Cloud. Autoscaling removes the need for the customer to constantly monitor the infrastructure and scale it up or down themselves. They can set limits to how their infrastructure scales and concentrate on the core business.

IaaS Delivery Models

The final piece to consider is how the IaaS is delivered. It can be delivered via a number of vehicles, which you are probably familiar with: Public Cloud, Private Cloud and Hybrid Cloud.
Public Cloud
Public cloud consists of shared resources deployed over the internet to the general public on a self-service basis. Windows Azure Services and IBM Blue Cloud are two examples of public Cloud.

Private Cloud
Private cloud is managed by the organization it serves. It may be an on or off-premises solution; but in either case, the server is dedicated to the organization and no sharing takes place. It can be argued that this removes many of the characteristics that makes a Cloud offering desirable.
Hybrid Cloud
Hybrid cloud is a combination of public and private cloud. It comes in two flavors: virtual private clouds and cloud bursting. Virtual private clouds exist when a public cloud sections off a portion of their own infrastructure for a single customer. Cloud bursting allows the customer to continue using their own infrastructure but moments of high demand get passed through their public cloud server. This allows them the best of both worlds – the continued use of their already purchased infrastructure combined with the scalability of public cloud.

In Summary

IaaS is a complex offering and expensive to implement, often prohibitively so for smaller businesses. It makes very little financial sense for your customers to build their own infrastructure when mature offerings already exist. That doesn’t mean they don’t need to understand IaaS. Understanding the building blocks is the only way to build the perfect solution to meet their specific needs. This is where you come in as a trusted advisor. You can educate them, advise them on the best solution, and provide the support that will allow them to focus on their core business, resulting in more efficient operations and higher revenue.

Contact Us

If you have additional questions or would like more information on Infrastructure as a Service through Arrow, contact or call 877.558.6677.

 Did You Miss Any of Our Other Articles in the A-Z Series?

Abstracted from Rackspace’s “Say Goodbye to DIY Data Centers

Last modified: March 27, 2019