The digital transformation is underway, with the cloud as its foundation and driving force of the technological disruption that enables agility and innovation at a lower cost. Accelerated by COVID-19, public cloud infrastructure spending has exceeded on-premises spending.
Cloud transformation is underway
Most companies have embarked on their cloud transformation, to develop the cloud native or migrate the legacy, but they are struggling to scale up and reap the benefits. This is due to the lack of transformation of the operational model, since it is impossible to reap the benefits of an agile and automated cloud infrastructure with processes that remain complex and manual.
The traditional production model is reaching its limits
Cycles that are too long. For an IT project, production occurs at the end and is often handled as a separate project, led by a production project manager. Its late involvement and alignment difficulties often lead to delays or trade-offs on test plans and quality.
Skills in technical silos. Technical expertise is organized in silos (server, storage, network, etc.) with widely dispersed tools, a lack of end-to-end transversality. ITIL models have often been misused, with requests often leading to a multiplication of tickets and round trips.
Highly manual operations. The practice is to make the settings by hand in the consoles, leading to errors and delays. The constant pressure on costs has sometimes led to offshore activities, which is no longer sufficient. Automation approaches are too scarce, based on in-house tools that are not very scalable and made inefficient due to dispersion.
Financial management giving little leverage on costs. A large proportion of infrastructure costs is generally almost constant – CAPEX, low-granular elements, or unused capacity. Above all, traditional infrastructures are accompanied by excessive service management, capacity, and numerous tools that significantly increase the price of infrastructures (markup of 30% to 200%!)
Four operational model transformation axes, aligned with the benefits of the cloud
1. Verticalization of production and product mode for agility and innovation
The main advantage of the cloud is agility and innovation at low cost. The aim is to improve the velocity and the time to market of the new features requested by the business. The cloud eliminates the slowness of the last-mile infrastructure and is therefore essential for agility.
To achieve this, the organizations must first be aligned in a logic of verticalization by business area to better accommodate the requirements and to improve collaboration.
With business products driven in scaled agile framework (SAFe), the Ops must work more closely with the agile teams, either by participating directly in the teams to process the requirements in plateau mode (rather than in ticket mode), or by participating in a system team that can be shared in a train (ART), or both in a pooling logic.
2. Infrastructure platforms for flexibility
Flexibility or scalability is also one of the advantages of the cloud expected by the business, in order to cope with activity variations (seasonal or related to the economic cycle) or merger-acquisition operations. The cloud brings this flexible capability but it also promotes the adoption of technical standards that increase flexibility.
This is the idea of platforms, invented by the automotive industry. Each cloud will thus be a platform with its standard services, managed as an end-to-end product intended for development teams and in a dynamic perspective, since the cloud catalog is enriched at high speed.
3. Strengthening automation with SRE and DevSecOps for resilience
More and more companies (and their CISOs) consider the cloud to be more resilient and more secure than the traditional infrastructure, due to its automation (patching in particular).
To benefit fully from this advantage, the engineering and automation aspect of operations must be increased, with DevSecOps and the role of SRE (site reliability engineer) popularized by Google, supporting the operational aspect of the products to which it contributes.
4. A new FinOps model for cost optimization
It is now widely recognized that a FinOps approach must be implemented to activate powerful cost optimization levers: reserved bodies, power off of non-prod environments, rightsizing, etc. However, it is not enough to set up a FinOps role and a tool, it is essential to involve the application teams, which must validate or implement these changes. The operational model must therefore be improved by adapting the financial processes, responsibilities and governance – since cloud costs are essentially direct costs, business IT becomes accountable for them.
A multi-speed model taking into account the existing situation
To meet the main challenge of the cloud, which is agility, the operational model must be adapted according to the type of applications to be managed, taking into account the legacy. A multi-speed model, adapted to the velocity of the applications and their architecture and able to change over time, must therefore be set up. It can be categorized as follows:
|Model||Optimized legacy||Cloud ready||Cloud native|
|Architecture||Strongly coupled||Tooled coupling||Decoupled|
|Organization||Business alignment (verticalization)||SRE team/system team||SRE integrated with teams + system team|
|Deployment tooling||CI/CD implementation||CI/CD with Infrastructure-as-code templates/automated tests||Blue/green or canary deployment|
|SLO (service level objective)||End-to-end (application availability, etc.)||Multidimensional (latency, query success rate, etc.)||Control by burn rate|
|Incident process||Constructive postmortem||Plateau mode management||Testing/simulation|
|Observability||Health map||Predictive systems/decision support||Proactive self-healing systems|
The future of the Infrastructure and Operations teams is at stake
The cloud is a profound transformation for I&O. Skills and activities must be developed, new approaches must be put in place to fulfil its promises. And above all, do not leave the product teams without Ops, since the cloud needs them to strengthen the culture of run, availability, and business continuity.
The I&O must reinvent itself, this transformation is not easy – but it is exciting!
To find out how Capgemini can help you get full value from cloud, download our Cloud Transformation: The Keys to Success whitepaper, or contact our team today.
Strategy manager, Capgemini