Skip to main content

This is really what it’s about for most firms: the cost.[/highlight]There are many ingredients to executing a reliable and scalable service. Reliability is a simple but useful term because it is a stand-in and abstraction for many of the related attributes of a good service: security, currency, fault-tolerance, redundancy — reliability is the end benefit whereas the related attributes are details. Each attribute is meant to mitigate the corresponding risk area.

The choice of whether or not to make use of a technology and/or outsource it to a provider often comes down to a cost comparison: do it in-house or outsource, and for which components. It is often a matter of whether a firm has — or can cost-effectively acquire — the skills, methods and materials to build and maintain a given service or a part of it. There are risks and benefits to both.

Risk mitigation of a vendor is also a cost, so don’t forget to enumerate the major ones like:

  • Platform or vendor lock-in;
  • Ability to migrate data in and out of a platform and vendor;
  • Conditions under which the cloud has gone down before, and the backup measures required;
  • Cost of downtime (often phrased as revenue and reputation earned/lost per hour);
  • Degree of reliance on the technology platform being put in the cloud (private, hybrid or public);
  • Disaster recovery and business continuity;
  • Quantifying risk as a measure of likelihood x dollar value impact to the business if it goes down. For you statisticians, this is the ‘expected value’ of a downtime event of a resource.
This is really what it’s about for most firms: the cost.

An outside provider can more often than not execute infrastructure very efficiently at large scales and then pass on the savings to their clients. The breakpoint can of course be a bit closer if you, the client, are already running a significantly sized infrastructure in-house and have already dedicated a large and skilled pool of human and capital resources dedicated to its construction and maintenance. Even in this case, the expertise of a third party provider cannot be discounted. In the very least, an audit of the infrastructure by a provider who has actually built similarly-sized (if not larger) infrastructure, can provide peace of mind. Consultancies may be able to provide great schematics and strategies, but a battle tested team can point out real-life scenarios and potholes that an otherwise purely theoretical handbook cannot.

There is simply no substitute for experience in this area. The technologies being used to deliver online services are constantly evolving, and there is absolutely no such thing as an invincible system. Knowing the benefits and risks of vendors and technologies come with on-the-ground experience.

The original question

All this would make anyone wonder about all the different things that can be “cloudified”. By now, you’ve likely realized that workloads like file storage, ERPs, billing systems, and even user desktops (known as virtual desktops) – pretty much anything – can be run online and accessible from potentially anywhere in the world. The workloads can be virtualized and hardware costs can be cut.

As new technologies being used to scale and grow businesses emerge, they need to be set up so that they’re reliable, safe, and properly maintained. The original question being discussed can really now be transformed into a three-part matter of: what problems and opportunities exist within your specific business, the workloads that can be run and your firm’s dependence on them, and the identity and structure of the business (as this will help determine who can help). Some workloads and business processes should be run on custom cloud infrastructure only internally and away from the outside world, whereas others can be wholly placed online.

Here are some examples illustrating use cases, the benefits, and the corresponding choice of cloud provider and setup:

  • Financial applications are an example of private-only and potentially only-on-premise cloud workloads. Uptime is crucial for such applications.
  • A corporate video conferencing system on the other hand need not be on premises. Public cloud providers with shared resources are often sufficient.
  • An inventory system available on tablets can be run on a private cloud, securely accessed over VPN by tablets. Secure mobility can massively scale a workforce’s ability to execute business processes.
  • Disaster recovery and business continuity processes and systems employ hybrid solutions, connecting local systems to those of the cloud provider. The underlying infrastructure must be minimally shared with other clients. Disasters are inevitable. Security and time to recovery must be tuned based on the organization’s individual needs. A private custom solution with industry standard cloud hybridization tools are required here.

Part 1: Cloud, what is it good for?

First part in the series make the concept of cloud crystal clear.

Part 2: Isn’t everything online part of the cloud?

This series deals with the concept of virtualization and the potential this idea has on business.

Part 3: Your identity dictates what you need

This series brings up the matter – probably the most important one in the discussion – of identifying internal needs and therefore who can help.