Introduction
In this post I share some of my observations of topics which have to be considered before making a decision about the future IT Platform design.
Like anything else within IT then there is no simple answer to whether an On-site Server Center, a Hosted Server Center or a Hybrid design for your IT Platform are to be preferred. The only thing for sure is that the majority of scenarios in the business case are identical when looking at the end-user experience, availability and security.
In this article I will pretend that licensing is not an issue as I intend to share some thoughts about using licensed vs. Open Source products and the licensing models in a later article.
When using the term; On-site I refer to the IT design, where everything are physically present on the geographical location of the office.
When using the term; Virtual I refer to the IT design, where several application servers are virtualized in one physical hardware box.
When using the term; Hosted I refer to the IT design, where the server centre are operated by a contractor and physically separated from the geographical location of the office.
When using the term; Cloud I refer to the IT design, where all the services (Application & Storage servers) are installed in a virtual environment managed and operated by a contractor.
When using the term; Hybrid I refer to the IT design, where the services are split between and handled both on-site and hosted.
It is my intention to highlight some of the issues I have seen being neglected in the Business Case and/ or by the management team.
The consequences have been overspent budgets and costly downtime during working hours – including a frustrated staff – as an “unforeseen” incident occurred.
There are more issues at stake
Whether you choose an On-site, a Hosted or a Cloud based IT Platform there are some identical scenarios, you will have to address in the business case.
The most common ones are related to performance, i.e. end-user experience to response time and availability.
What happens if your primary application server crashes due to a bad hard drive or an upgrade fails?
In the scenario where you have a primary and a secondary environment with a failover configuration, then scenarios related to the physical location and distance will have to be included in the evaluation.
Have you ever thought of the power supply configuration?
One test I have never seen being executed is a simulation of a human error like pull out the power cable in the server and then plug it in again or test the emergency button next to the door and then within 10 seconds turn on the power again. British Airways had an incident in April 2017 with some financial impact due to a human error scenario nobody paid attention to.
Another important issue when evaluating the future IT Platform design is related to the physical location of the data stored in the Hosted or Cloud design. Depending on the country of residence then there are some legal compliance to where and how data have to be stored.
As the business focus is primary on cost, i.e. hardware, flexibility and staff, the owner of the Business Case has a tendency to do as Lord Nelson under the Battle of Copenhagen 1801, but we are now in a situation, where all aspects needs to be included in the business case.
Finally, the organization maintaining the future IT platform needs to be addressed with respect to roles and competences. I recommend here to include a section related to training of existing staff as they have valuable knowledge for the business.
Business expectations
Business teams expects IT to be a service available 24/7 – including an availability to handle any requests (new/ replacement/ change/ etc.) in a rapid pace.
The requests of today includes tools for sharing information, collaboration, analysis and availability.
The business teams have increased the use of mobile devices in their daily operations. Introducing tablets and smartphones forces a different approach to the IT design.
Since it became common to allow the employees to use their own devices we have a new dimension within the IT Platform design that requires a second thought in terms of security.
The customers wants the availability of self service, i.e. access to product documentation and/ or ordering. This service requires full attention to the security, i.e. classification of data, threats and so forth.
On-site scenarios are not equal Hosted scenarios
The Business Case for moving the IT Platform engine room from an On-site to a Hosted solution should include proper risk analysis of the potential scenarios related to having the company IT engine room at another geographical location.
Many of the Business Cases I have seen simply states that there is no difference in the risks between having the engine room On-site vs. Hosted, which I find to be an understatement.
I have over the years experienced downtime due to such a simple incident as a broken cable connection between the office building and the IT engine room – despite the design included a fail-over scenario in order to secure a 99% uptime including the monthly 6 hrs. service window.
A simple investigation of the actual location of the physical cable trunks would have revealed, that both the primary and the secondary cable happened to be in the same trunk.
A hardware limitation in the IT platform design
To simplify the explanation, then each business system requires a physical server in order to secure a stabile performance. When an organization increased the use of the business system it often resulted in replacing the server with a new and better performing server – not always reserved in the yearly IT budget.
Using one server per system add up requirements and additional costs to the server room, i.e. power supply, cooling capacity and physical space. In some situations these requirements have limited the option to add/ upgrade existing hardware, especially if the outdated servers still are in use for other purposes, i.e. a development server, a test server or for another system.
The development within hardware designs, i.e. Blade Servers did open up for a better use of the physical space within the server room.
Introducing the Virtual design opened up for a more flexible use of the hardware components, especially when using Blade servers.
By changing to a Virtual design using the latest generations of hardware gives in my opinion the option to continue using the existing server room and yet double the capacity compared to the initial IT platform.
Virtual design
The Virtual design gave the IT department the opportunity to establish extra servers for testing and training purposes, i.e. easily create a copy of the productions environment when required.
The only limitation here would be related to sizing and performance.
Such flexibility requires extra staff in order to keep track of versions, clean up, etc. – activities which often happens to be underestimated both related to staff and time.
Cloud design
Introducing the services offered by the Cloud design have moved the engine room away from the IT team to a contractor.
This has been seen as an advantage related to service, cost and capacity – you “only” pay for what you actually are using as the technical support are handled by the contractor.
Changing the setup to a Cloud based design works very fine if the business are using a “simple” design, i.e. few applications, a few interfaces to external systems and only doing business within the region.
The challenges comes when the business expects a dynamic handling of their IT requirements.
Using the Cloud design requires a “Contract Control Centre” for managing change requests as there will be a financial impact no matter how small the change will be.
Technically then Cloud designs are build on Virtual servers so changes can be implemented quickly, but you remain responsible for the compliance within all the requirements.
The hybrid design
The Hybrid design can be used in the scenario where the company chose to keep some critical business systems on-site and only uses the Cloud services for noncritical business systems.
Another scenario will be where the company uses Cloud services for storage, backup or disaster recovery.
The simple technical difference between On-site and Hosted/ Cloud design
In short the only technical difference is the available hardware and the awareness to cost.
The internal IT department are limited to what has been approved previously, where a contractor due to their business strategy will have enough spare capacity available for expansion.
Having the IT platform on-site there was(?) a tendency within the organization to forget that any change has a financial impact.
Using a contractor generates an awareness of that there are a cost connected to a change request – if communicated.
I would recommend a Change Request process which also incorporates the financial impact to the budgets when implemented.
The IT platform design is complicated
This article was published initially on LinkedIn on 7 September 2017 (Some modification of the wording has been executed in this version).
Image Credits:
| View from the sky | Photo by NASA on Unsplash |
| Server Room | Photo by Ian Battaglie on Unsplash |
| The Game | Photo by Michał Parzuchowski on Unsplash |
| Excavator | Photo by Kuan Fang on Unsplash |
| Back Plane | Photo by Massimo Botturi on Unsplash |
Thank you for having read this article – hope you have enjoyed it and that it has given you some ideas of where to start improving your own business or individual role, when it comes to the use of IT.
Best wishes for the future.