This article is based on an online roundtable titled "The (in)convenient facts about the cloud,” organised by Pointury on August 24, 2022.
On Wednesday, August 22, we engaged in an open discussion facilitated by Peter HJ Van Eijk, Associate Professor of Cyber Security & Cloud at Utrecht University of Applied Sciences, and Bart Schillebeeks, Senior Enterprise Architect for Cloud & Infrastructure. During this event, a group of CIOs gathered to share their experiences with cloud environments.
Cloud environments have gained immense popularity due to their ability to offer rapid delivery, high availability, extensive functionality, scalability, and flexibility. Even companies heavily reliant on SAP are transitioning to Azure Cloud and other cloud platforms as a result of digitisation projects. Some organisations have entirely migrated from on-premise physical and virtual servers to cloud environments. In particular, start-ups are forgoing the creation of their own data centers.
The three primary cloud providers have distinct specializations:
- AWS focuses on business-to-business applications.
- Azure is geared towards MS Office integration.
- Google Cloud centers around big data, extensive search capabilities, and artificial intelligence.
The reality is that over 80% of organizations operate in a multi-cloud environment. Many of these organizations have a primary cloud provider (such as Azure) and an additional environment for specific purposes (like using Google Cloud for big data applications).
This multi-cloud approach can be attributed to several reasons:
- Organizations aim to leverage the specializations, unique applications, and tools offered by each cloud environment.
- Different divisions globally making choices for diverse cloud platforms can lead to a multi-cloud setup. Mergers, local provider popularity, and the availability of experts can contribute to this.
- Even shadow IT can introduce unplanned SaaS solutions in different cloud environments.
Compared to traditional on-premise data centers, cloud environments require fewer specialized experts due to the abundance of user-friendly tools.
Can one switch cloud providers?
Yes, switching is possible with preparation. Some companies have even transitioned from Microsoft 365 on Azure to Google Workspace in the Google Cloud. You can also opt for cloud-native applications to ensure your organization's independence from specific cloud environments and their native applications. F5 provides services that work across various cloud platforms, eliminating the need to manage each environment individually and enabling application distribution across multiple clouds.
While standardization might seem advantageous for switching between cloud providers, the complexity arises when dealing with big data in Google Cloud, virtual machines in AWS, and containers in Azure. Standardization may primarily involve governance, as trying to standardize services across diverse cloud environments might negate the advantages unique to each environment.
It's essential to establish from the outset that applications function across different cloud environments.
How can a solid security architecture be built?
Security remains primarily your responsibility, particularly in an IaaS solution. While cloud providers safeguard their services, they won't shield against distributed denial-of-service (DDoS) attacks on your applications. It's also your responsibility to implement cybersecurity with minimal resource consumption. Key starting components include "identity and access management," "login," and "cloud composition and compliance management."
Is there a business case?
Creating a new environment begins with analyzing use cases and building a business case. The process concludes by implementing financial operations procedures.
How can costs be minimized?
Cost savings can be achieved by shutting down non-critical services during nights and weekends. Similarly, utilizing "AWS spot instances" allows running non-critical applications during low-cost periods. Comparing global service proposals from different regional offices of the same provider can also help reduce costs.
Is cost allocation easier?
The initial technique for allocating costs to business units involves using "projects" in Google Cloud, "accounts" in AWS, and "subscriptions" in Azure. A more advanced method employs tagging and labeling to assign costs to different business departments.
Where can required expertise be found?
Existing infrastructure teams require limited training to become proficient in cloud environments, depending on the chosen platform. A Microsoft engineer will naturally adapt to an Azure cloud environment.
The era of on-premise solutions seems to be fading. Even in this context, it's advisable to prioritize new developments as candidates for cloud deployment before migrating legacy systems.
We extend our gratitude to all the speakers, including CGI and F5, as well as all the attendees, for their valuable insights and contributions during this enriching event.
