Based on my amazing cloud experience, here are a few reasons why
|Reason||My personal on premise experience||My cloud experience|
|Agility and cost benefits||Slow and costly server builds||
Build servers within minutes, pay per use.
Agility is number one benefit, esp. for dev/test environments time is really the key. Standing up new environments, copying, scaling it, and tearing it down when not needed in just a few clicks of a button. Technology just works!
|Elasticity and reliability||issues relating to poor system performance, server(s) and network overload||
Scalable network (pay extra for higher IOPS). Scale instances as required (vertical or horizontal)
We are not limited to what we’ve got, we have almost unlimited options to improve performance. Elasticity is definitely a bonus; we can scale systems quite easily and virtually no capacity limits (i.e. hardware limit esp. disk). The word capacity seemed redundant or no longer applicable in that kind of environment
Reliability – not a single incident (downtime or degradation of service) since we used AWS. Customers often realise the dev/test environments are sometimes more reliable than the on premise ones, even against production.
|Capacity, cost transparency, predictability||Capacity issues, the common “we are running out of disks”, it is really hard to predict||
Virtually no limits, resource is always available (for most customer usage). All costs are visible, tools are available to reduce cost
Visibility – Cost is something that needs to be managed; the fact that customers have a full visibility of what they are paying even up to the last cent then it does really take the burden of uncertainty. Cost then becomes secondary, and often customers rather eliminate uncertainty. AWS also provides tools for trend analysis (and full report of what they’re paying for, even has tools to advise how to reduce cost!). We have all visibility of how much and where we are spending our money. In reality though, as long as we apply a lot of smarts in our cloud expenditure, we are certainly paying a lot less money than compared to on-premise. We can even run a full production system, complete with DR/failover capability for only less than $50 a day!
|Reliability and durability||Backup failures, untested data retention and durability||
Not a single backup failure, assurance of eleven 9’s or nine 9’s of durability (up to you)
A painpoint in BAU, in AWS we never had a single failure. We set these up once, and we know it just works!
|Agility and risk management||Very long, costly and high risk patching cycles||
Use latest AWS template, replace OS, try and fail a lot without impact until it becomes perfect the time you patch production
Agility and risk management – something we don’t normally see in BAU, we even have the ability to achieve near zero downtime when patching
|Redundancy||Costly HA and DR solution||You just have to avail and configure HA and DR solutions. Cost is based on amount of resource you use. If you terminate the systems (e.g. after DR test) you wont be paying anything at all|
In as much as our customers want the pains in supporting SAP systems to go away, we as consultants want those gone (trust me)
Service providers like us and our customers should be working together on value adding and exciting new work.
There is nothing wrong with ITIL processes (incident, problem, change managemet) but come to think of it, should we focus on these processes in our organisations or should we rather shift our time and effort to things that really matter to the business?
And lastly CAPEX to OPEX is still very important point. By spending more on services, then we gain a huge room for innovation. We should leave the datacentre mentality. Leave the heavy lifting to cloud providers who know exactly how to do their job 😉
* infrastructure as a service like AWS