Once upon a time there was an IT organization that said “We’re going to solve all our problems and save money: we’re going to the cloud! And there they went, a new adventure ahead. It would result in significant savings in hardware and management costs. The new supplier would take over everything, so there would only be one point of contact: ideal! The first users were migrated and happy: it worked really nice and fast, a fresh environment, and finally the new Office 365, nice! The vibe was good, so after the holidays the rest of the users were migrated too. But then: in September some users started to complain: it wasn’t always that fast. The supplier still says it complies with the SLA and no one knows exactly how slow it actually is. What to do? We see this scenario at our customers on a regular basis and it has a number of causes. We discuss the four most common ones.
1. Fragmentation of responsibilities
Each his own profession, so the workspace environment is purchased from supplier A, who rents server capacity from supplier B. The backend applications are managed by party C, who uses internal hosting from department D. All parties take responsibility right to the front door: whether their product works well within the total IT solution is not part of the SLA. In the event of problems, suppliers often point at each other and nobody feels responsible for the functioning of the whole. Meanwhile, the customer assumes that by outsourcing his environment, problem solving is also well invested. Our advice: make sure to stay in control by measuring how your suppliers perform form the beginning.
2. Shared environments
If the brochure states that a supplier can ‘work efficiently’, he probably saves costs by purchasing large quantities and delivering them to several parties. The result is just like in a cheap hotel: the water from the tap is nice and hot, but not between 7:00 and 9:00 in the morning when everyone wants to take a shower and share the same water supply. For example, the transition to the cloud often means nothing more than ‘running on other people’s servers’, which means that old problems simply persist or new delays occur. Our advice: be realistic in the cost savings that can be achieved.
3. ‘Taking control’ means migrating
When a company has grown out of its own environment, a new supplier will offer to ‘take over’ the environment. That sounds good and often works well. For a moment. The willingness of suppliers to solve problems in an environment foreign to them, is often limited so an ‘alternative’ solution will soon be proposed. Existing problems are still not solved and it is proposed to switch to a ‘standard environment’. This means a migration from the old infrastructure to that of the new supplier, of course with new conditions and budget. Our advice: think about this in advance and make agreements with the supplier.
4. Unforeseen consequences of services
Cloud providers often offer an agreement with standard services in it: tempting options of which the consequences are not always clear. For example:
- “Your documents quickly available” means: documents travel with you and are stored on the nearest server to guarantee the agreed speed. If you are on your way to your holiday destination in Vietnam and need to quickly edit a document, it will be stored on a server in Singapore, under Singapore’s privacy laws. Is this still in accordance with your wishes and policy?
- “Unlimited return to each previous version” means: the version history is kept, so the more changes are made, the slower it gets. After two years of intensive maintenance on shared source documents, working with these efficiently probably becomes increasingly difficult.
- “Desktop environment is remembered, wherever you work” means: the layout of the work environment is stored in a personal profile. Nobody dares to restrict this, so after a year, a user carries so much history with him that logging in and out can take up to 10 minutes.
Our advice: ask for more information and record agreements about performance, for example in an eXperience Level Agreement (XLA).
Should you go to the cloud? That is up to you to decide. Our message here is : learn from the pitfalls of others and make clear, measurable agreements with your cloud supplier about the availability and performance of services. And keep control of it yourself!
During his career, Marcel has helped many organizations such as the NS, ProRail, ASR and various municipalities to improve their IT services. Speed, capacity, scalability and stability of the software are the key to success. Marcel started his career in 1996 as a developer, followed by activities such as performance testing, load testing, performance troubleshooting and in recent years automated performance testing in Agile development projects. Within Ymor, Marcel is known for his expertise in Dynatrace, troubleshooting and automated testing.
We would like to keep in touch! Join our newsletter and be informed about new developments, blogs, events and more.