5 key components for your Cloud Platform
After building Cloud Platforms for numerous organisations, I wanted to compile a list of the key components that work together to accelerate your teams.
Some of these implementations were brown brownfield and some greenfield. This factor does not change the components involved in my view, just the manner in which you roll out these components.
Skipping any of these components will drastically slow you down on your journey for reasons I have tried to include with the sections of the components themselves.
Worth noting that these components also don't differ depending on your Cloud Service Provider, although some CSPs will include them with their Landing Zone or Foundation Platform implementations.
Multi-Account
The ability to create multiple accounts within your Cloud organisation that are identical without manual intervention is integral to being able to scale your company in the Cloud.
Having previously tried to deliver a Cloud platform in a single account with only network-level segregation, I can vouch for the fact you do not want to do this.
Sure there are strategies to enable your Company to move faster than you can move on-prem like this, but you will not get the full benefits of being in the Cloud, use account-level segregation to reduce blast radius and speed up the delivery. ?
Authentication
Most of the Cloud Service Providers will enable you to handle your authorisation via IAM. But one of the most important services to build or leverage first of all is a robust authentication service, that can be used by all of the services that are consumed by your engineers.
The end goal here is to have the same database of usernames and passwords for every service within the estate, giving the end result of Single Sign-On to your engineers.
If you don't solve this upfront and use different authentication services throughout the organisation your engineers will be wasting time remembering different passwords and all the cognitive load will ultimately slow down your development cycles.
Also ending up with the different options within the estate will cause confusion, not to mention the different maintenance costs associated with them all.
领英推荐
Telemetry Data
Telemetry data refers to metrics, logs and tracing of your applications. Having a centralised service in which all employees can view all of the telemetry data that are important to them to be able to do their job will accelerate your delivery massively.
If your engineers have to search around in different locations for this data from different systems, this wastes their time while debugging problems.
This can cost you money while your systems are down, of course, but also wasting time in debugging means that you are not adding new functionality or fixing the bugs so again it is costing you money.
Continuous Integration
This is very important to be able to ship code rapidly, having a good process in this area.
The best option here, at least in the beginning, is to go for something hosted, so that your teams don't need to worry about the infrastructure.
Yes, you will be charged for usage and yes this cost would go away if you owned the infrastructure. But why do you want to waste your company resources ( money and time ) worrying about the complexity of this infrastructure when you can pay your cloud provider pennies to look after that for you.
Believe me, as much as you want them to be, your internal teams are not experts in how to deliver a CI service. Trust the experts. You can refactor later when teams are moving at speed and the cost of ownership needs to be lowered. However, I am very confident that with the proper business analysis, you will find the cost of implementing it yourself to be wasted money.
Security Baseline
Increasing your visibility of the platform and applications compliance with your security policy is massively important when building out your foundation platform.
Making sure that with every build you are aware of how much closer you are to compliance, or whether you have regressed against compliance is motivating for your engineers. It should also help to lower breaches based on security vulnerabilities or holes that your team should know about.?
On top of that visualising and socialising the level of compliance can really help with getting things over the line more quickly. This is because multiple teams can agree on the state of play and allow certain exceptions to the policy for short time frames, this practice happens anyway, but is much easier to navigate and therefore saves money when the level of compliance is easily viewed.
CEO and Owner at LVIVITY | Tech Entrepreneur | Experienced guide in your software development journey
1 年Carl, thanks for sharing!
COO | Quema | Building scalable and secure IT infrastructures and allocating dedicated DevOps engineers from our team
1 年Carl, thanks for sharing!
Data & Digital Architect | Consultant
1 年Carl, thanks for sharing!