Managing Costs with FinOps

Building and running a data center in public clouds can be more cost-effective than going the traditional way. However, running a data center in public clouds still costs money. Hence, as ever, you still need ways to control costs. financial operations (FinOps)—is all about cost control.

This chapter focuses on the starting point for managing FinOps in multi-cloud environments: the provisioning of resources and the costs that come with the deployment of resources. We will learn how to keep track of these costs in the public clouds of Azure, AWS, GCP, Alibaba Cloud, and Oracle Cloud Infrastructure (OCI). However, we start with a brief introduction to the principles of FinOps.

In this chapter, we’re going to cover the following topics:

• Understanding the principles of FinOps
• Defining guidelines for provisioning cloud resources
• Defining cost policies for provisioning
• Understanding account hierarchy
...

Join our book community on Discord

https://packt.link/multicloud

Market researchers such as Gartner expect that in 2030 over 30 billion of devices will be connected through the Internet of Things (IoT). It would mean a skyrocketing growth from the current number of connected devices: at time of writing the number of devices count around 14 billion, still an incredible figure. We have connected devices in our homes, in factories and even in cars. All these devices produce and receive data.

The challenge that companies face is how they can monitor and manage these devices, preferably from one platform. Cloud providers offer centralized IoT platforms as a solution to this challenge. In this chapter we will study the architecture principles of an IoT ecosystem and discuss how cloud can help in managing IoT devices. We will explore some of these cloud solutions and also look at crucial elements in IoT, such as connectivity and security.

We will cover the following topics in this chapter...

Choosing the right platform for IoT

There’s a lot of talk about the Internet of Things, the internet that connects everything: every device, perhaps even every human. Let’s first define a ‘thing’. Basically, a thing can be anything but typically we talk about devices that have a connection to a network through, for example, Bluetooth or Wifi and through this connection to the internet. A device can be a machine or someone carrying a device with a Unique Identifier (UID). The device is capable of autonomously transferring data over a network, so without human interference. In literatus this is referred to as machine-to-machine (M2M).

A thing can thus also be a device that is ‘attached’ to a person. You can think of implants that monitor the health status of patients. But we also think of cars with all sorts of sensors alerting the driver when something is wrong with the car and already sends a message to the dealer to get an appointment to fix...

Monitoring IoT ecosystems

The most important challenge that architects should address in IoT is, not surprisingly, security. The security estate of IoT devices must be monitored continuously. To understand the risks better, we can have a look at top risks that are listed by OWASP, the Open Web Application Security Project. We only list the top five.

• Weak passwords
• Poorly protected network services
• Poorly secured interfaces
• lack of update mechanisms for security rules and patches
• Use of outdated components

The top risk however is the lack of device management and leaving the data transfer unmonitored. However, this starts with knowing where the devices are. But knowing where devices sit is not sufficient; we must also know what these devices are doing, what sort of data the devices collect and how they collect it. Observability is the key principle that any IoT architecture must comply with.

This is already the biggest challenge in IoT. Think of the fact that an IoT ecosystem might consist...

Designing for connectivity to cloud

Before we get to monitoring and managing an IoT ecosystem, there’s one crucial step that we must take first: connecting the IoT devices to our cloud. Gateways and edge computing are solutions to this.

We can’t simply have thousands of IoT devices connect one-by-one to systems in the cloud. First, every connection on itself is imposing a risk of intrusion and with that the risk of a security breach. The solution to this is to have all connections targeting one system, machine or instance that sits in front of the cloud environments, before data of IoT devices is entered to systems that are hosted inside our cloud.

The IoT gateway is such as system. It can be a virtual machine or a service, but in all cases, it serves as the connection point between the cloud and the IoT devices. All data that flows between devices and the cloud environments, will have to pass the gateway. The IoT gateway is the central connectivity and data controller....

Connecting IoT with IPv6, LoRa and 5G

In the previous section we discussed how we can connect IoT ecosystems to public cloud environments through gateways and edge computing. It’s important to understand that IoT used different protocols to communicate amongst each other and eventually to the cloud. The reason for this is the massive number of devices that must be connected and with that the amount of data that is transferred from these devices to the cloud. The infrastructure must be capable of absorbing this, but it requires different means of communication.

We need connectivity that is able to connect machines with machines and continuously transport real-time, small chunks of data between the devices and cloud environments.

There are several emerging IoT standards, including the following:

• IPv6 over Low-Power Wireless Personal Area Networks (6LoWPAN): the Internet Engineering Task Force (IETF) defined this as an open standard for connecting and communication between low-power...

Summary

In this chapter, we discussed the basic architecture principles for Internet of Things-ecosystems (IoT). It’s fair to say that IoT architecture is extremely complex, exposing a lot of challenges. The first challenge is the scale of an IoT ecosystem that can easily hold hundreds to thousands of devices. All these devices must be secured and monitored. Next, these devices send data that must be analysed. This requires a centralized platform.

We explored the IoT solutions of Azure, AWS, GCP and Alibaba and what they have to offer. The services vary from update management of devices to security monitoring to detect and respond to potential vulnerabilities in IoT ecosystems. An import element in IoT architecture is connectivity. In the final sections we studied the connectivity to cloud using IoT gateways and edge appliances and the various protocols to enable low-powered devices to connect and communicate with each other. Last, we explored the possibilities of 5G and how this...

Questions

1. What AWS solution would you use to build, configure, and deploy applications to IoT devices?
2. Both Azure and AWS offer appliances that can be used for edge computing. Name the two solutions.
3. What does LoRaWAN stand for?

Further reading

• IoT and Edge Computing for Architects, by Perry Lee, PacktPub, 2020

Summary

This chapter started with a brief overview of the principles for FinOps: financial operations in the cloud or cloud financial management. We studied how we can provision resources for various clouds and then learned how we can track costs that are related to these resources. Before we can track resources, view the associated costs, and validate invoices, we must understand how cost management works in the cloud. We discussed the cost tools in Azure, AWS, GCP, Alibaba Cloud, and OCI. All these providers offer comprehensive toolsets to provision and identify resources from their respective management consoles. However, we must understand some general principles, such as license agreements and tagging.

In this chapter, we discussed the foundation of FinOps. In the next chapter, we will elaborate on how organizations can implement and develop cloud financial management, including the setup of a FinOps practice, using the FinOps maturity model.

Questions

1. If we want to run a trial period in a public cloud, what type of agreement would fit our needs?
2. Cloud providers use different technology to provision resources. What technology do both Alibaba Cloud and OCI use?
3. What is the discount program for large accounts in AWS called?
4. True or false: the pricing calculators of cloud providers are free to use.

Further reading

• The Road to Azure Cost Governance, by Paola E. Annis and Giuliano Caglio, 2022, Packt Publishing
• AWS FinOps Simplified, by Peter Chung, 2022, Packt Publishing