Cloud Economics - Consulting Point of View
Introduction
Most organizations currently use one of the major hyperscale cloud vendors as a primary solution for their infrastructure needs. Enterprises adopt some sort of cloud strategy - either a cloud first strategy or defined cloud workload strategy. As the maturity of the cloud solutions have increased, it is becoming clearer that hybrid and multi cloud solutions are the current norm and solutions have to fit into this multi-vendor, multi connected model.
There is also the existing Data Center investments that need to be factored. Most enterprises also have private clouds set up and integrate with the hyper scale vendors and try to achieve a unified system of management.
Continual changes and new services being provided everyday indicate the architecture to be modular, agile and dynamic. While this is great news for the business as they can achieve the time to market rapidly, IT owners need to manage the costs and ensure cost overruns are managed effectively.
The public cloud is an extremely complex computing model and business can be sure to struggle with budgets and costs unless proper plans, procedures and processes are put in place.
Business Challenges that we are observing
Currently among multiple customers that we have been involved in cloud journeys and transformations there are common threads of the challenges that have been observed.
At the top of the list is rapidly spiraling costs of operations as cloud technologies offer more and more services that businesses adopt and try to accelerate their outcomes. The last few years have been about organizations moving their workloads to cloud in a usually lift and shift fashion.
While this moved the data center to a service provider, it really did not take the cloud’s computing model into design of applications and outcomes. While this was in line with the expected strategy, and also lent some sense of pricing in line with the old schemes this resulted in support and release management issues thereby pushing IT to adopt more cloud native solutions and development processes.
Also a major challenge has been into a cloud vendor lock down and investing resources – human and technical with any one of the leading vendors. While the end outcome may still be the same, complexities arise rapidly as there is no single standard across vendors which in turn could result in cost overruns.
Another Major issue that we observe among enterprises is the decision making process does not clearly translate to everyone in the system as to which vendor the enterprise is going to sign up with. Technologist and geeks try to make everything work and pick the best across different vendors. The overall strategy if not shared with the operations, the businesses and the architects can effectively kill the projects and also can guarantee cost overruns.
Multi Cloud strategies need proper analysis and details before execution as this can be extremely useful in cost controls
There still is the legacy mindset of moving the machines alone to the cloud – this in effect does not offer much. Many a time we have seen machines consuming .001% of the allocated memory and CPU and clients paying for complete servers. A mindset change is required to adopt functions, lambda and other services that offer a much better cost to utilization factor.
What our clients are asking
Some of the typical questions that we are facing today are as follows:
· How to effectively manage cloud costs?
· Is there a standard on cloud expense management ?
· How do we manage costs better?
· Do we really need all these services that we are paying for?
· Can we move our workloads dynamically across vendors with no downtime?
Addressing the above, the answers to the above is not entirely straight forward since there it is extremely critical to understand the business layout as to why something has to be set the way it is and how effectively it can be re imagined.
The following guidelines will help
a. A sustained and aggressive monitoring of resources that are being consumed to what was originally thought of. It is important to monitor the usage week on week, month on month till a pattern emerges
b. Are you using the hyperscale cloud as a DC or is it a computing model? A careful and in depth technical analysis will help drive the outcome for this. Both have to be dealt separately and the cost for a complex service may appear high, but may be much cheaper when compared to investing time and person into setting up internally on premise
c. Very critical to have an agile and modular architecture with the ability to interchange cloud vendors easily. End of the day, the outcome of any Digital strategy is to help businesses achieve more
d. A rigorous governance needs to be in place with business owners, technology team members and account managers from cloud vendors to determine if the cost curve is per plans and what can be done to continually improve this.
e. Have a framework and pre-built configurations and templates for small, med, large and complex services.
f. Train everyone to become digital … this is easier said than done, but the world is evolving this way , the way most people think in terms of servers and storage, the next generation is going to be thinking in terms of services, packages and automation
In summary, the above are not great thought patterns, however they work. This is because end of the day, any solution has to conform to (1) the law of physics (2) the law of economics and (3) the law of the land.
My point of View
·
Need of sustained and aggressive monitoring of resources
· Look at Cloud – As a computing model or as a DC?
· Review Architecture
· Review Governance
Comments