At its best, the cloud can deliver high value solutions in a cost-effective way, but it can also act as a financial black hole. As a customer, you have to navigate safely between these extremes and make fact-based decisions. It turns out that the suppliers often act as single-minded technology optimists.
An impossible decision
A while ago we had the opportunity to help a client prepare a basis of decision to be presented for the company’s senior management. As a result of changing times in their industry, they experience that the organization is becoming more distributed and autonomous. Previously, they had a large headquarters and a couple of small branch offices, now they have a presence in a larger number of locations spread across the planet. At the same time the need for centralized services decreases; they manage with fewer employees at head office. As a result of this, they will shortly be moving out of their large head office, in favor of more modern premises with barely half of the current space.
IMPORTANT NOTE
This blog article is approved by the client. To protect their anonymity, we have changed, or omitted, data that could lead to identifying them. The changes, however, are not of such a nature that it gives an unbalanced comparison of the solutions and Public Cloud platforms described
In an ideal world, it’s obvious that Public Cloud is the smart solution for a business with a growing global presence, but our client, like so many others, stands with one foot in the past and one in the future. They have moved standard applications such as Office, email, document archive and some business applications to SaaS, but are still left with 130 virtual servers running in their on-premises data center. Around half of these represent business-critical systems, while the rest are present because neither IT nor the business have been able to allocate the time needed to consolidate and remediate old systems.
A round with existing IT suppliers contributed to further confirmation of Public Cloud as the right choice. All of them recommended Microsoft Azure, with slightly different approaches to the migration process itself. The IT manager felt that this was not an adequate basis for making a good decision and contacted us for assistance.
Facts or hope and fear?
Facts and real alternatives can be surprisingly scarce in decision-making processes. Instead, the basis for decision-making is filled with hope and fear: the outcome we hope for, versus the outcome we fear. After all, this is in the nature of the future, we have little or no facts about it. At the same time, all decisions concern the future.
At Finalyze, we have 5 rules that are always followed when building a foundation for good decisions:
- Limit the scope
- Set the time horizon as short as possible (3 to 5 years = 3 years, 1 to 3 years = 1 year)
- Find the real alternatives (alternative route to destination, not just another vehicle on the same road)
- Find as many disadvantages as advantages for each alternative
- Use probability calculations to separate what is close to facts from hope and fear
We only spent a short workshop on complying with the first 3 rules – we came up with the following as a starting point for further analysis:
130 virtual servers of varying value to the business, running in local data center. The servers are to be reduced in number as quickly as possible. Some of this can be carried out before the move, but most of the consolidation can be carried out over 3 years. After the consolidation, there will remain around 40 servers.
| Solutions | Migration alternatives | Technology alternatives |
|---|---|---|
| Public Cloud | Migrate first, then consolidate | Azure, AWS, GCP |
| Consolidate first, then migrate | Azure, AWS, GCP | |
| Private Cloud | Migrate first, then consolidate | Nutanix, VMware |
| Consolidate first, then migrate | Nutanix, VMware |
- This decision is not about creating new value for the business, but about maintaining and continuing existing functions. Cost and risk should therefore be given primary importance.
- None of the alternatives should be an obstacle to future innovation and digital transformation.
- The Private Cloud option must be full OPEX and not require capital investments in hardware, licenses or data center.
The Economic Analysis
Decision-making bases like this do not need calculations carried out with millimeter precision. All experience suggests that you will not hit the nail clean on the head with your calculations anyway. If the process is carried out correctly, and in the end you are still left with two options that are impossible to choose between based on the facts, then you are allowed to use your gut feeling – the future is never an exact science.
We made use of the platforms’ own migration and calculation tools to estimate the price and used the discount models that came out best for the price in each case. We also compared our calculations with the estimates that the client had already received from existing suppliers, where possible. For the Private Cloud solution, we chose to obtain prices for one of the technology options in the first instance. This should prove to be sufficient to make a decision. For this alternative, we also had to price in data center rental, installation of hardware and rental of redundant power and redundant internet.
We prepared 3 scenarios:
- Worst Case: That the client still has 130 virtual servers running after 3 years
- Best Case: That the client manages to deliver upon promises regarding digital transformation, and only 40 servers are left after 3 years
- Most Likely: They end up somewhere in between Worst Case and Best Case.
In the charts below you can see a 3-year comparison between three Public Cloud and one Private Cloud solution. The first diagram compares the three scenarios for each solution, while the second diagram shows the financial uncertainty associated with each solution
The result
In this case, the decision was easy. The Private Cloud solution represented a significantly lower cost while the financial risk also was lower. In addition, the analysis showed that the operational risk was somewhat lower for this solution. This does not mean that Private Cloud is always the right choice. The number of virtual servers, the load and the time perspective are the most important factors when deciding this. If the customer had 10 virtual servers, or the timeframe was one year, Private Cloud would probably not have been a part of this analysis.
What is worth noting for everyone is the price differences of the 3 major cloud platforms. The fact that GCP (Google Cloud Platform) is so far below in price here is mainly due to two factors:
- GCP has more favorable pricing on virtual machines that require a lot of processing power. In this case, 15-20 resource-intensive servers accounted for a disproportionately large part of the total price across all 3 platforms, this became a clear differentiator for GCP.
- The discounts for 1-year and 3-year reserved instances also turned out particularly well for GCP.
Microsoft is the dominant player in Norway. A majority of the competence, with both customers and suppliers, is on the Azure platform. AWS is a good way behind, but has a certain prevalence, we are of the opinion that more and more of the large suppliers in the Norwegian market are also building competence on this platform. GCP is smallest, both global and in a Scandinavian context, therefore it is relatively rare for Norwegian businesses to receive offers for GCP.
However, there are no technological reasons why one should not choose GCP or AWS to run both virtual servers and many other workloads. It is possible to create equally secure and robust infrastructure solutions on all three platforms, as long as you have access to the right expertise.
English 
Norsk bokmål