example deployments, more to be added of course

content/prologue/deployments/index.md

@@ -1,7 +1,7 @@
 ---
 title: "Deployments"
-description: "Discover the different deployments and use cases, learn more about the architecture."
-lead: "Discover the different deployments and use cases, learn more about the architecture."
+description: "Kerberos.io deployments and examples"
+lead: "Kerberos.io deployments and examples"
 date: 2020-10-06T08:48:57+00:00
 lastmod: 2020-10-06T08:48:57+00:00
 draft: false
@@ -13,36 +13,43 @@ weight: 103
 toc: true
 ---
 
-In this section we will explain some use cases and best practices, which might help you to define the architecture that fits your needs.
+There is no good or bad deployment. All the architectual decisions you make, should focus on the goals of your project and what you or your customer would like to achieve.
 
-> Please read the [introduction page](/prologue/introduction/) first, before moving forward.
+In this section we will explain the most common use cases and best practices, which might help you to define the architecture that fit your needs.
 
-When starting with Kerberos you might ask yourself, how should I deploy these so called, Kerberos Agents? Well, there are different options. You can install a Kerberos Agent on a Raspberry Pi, VM, Server or even in a Kubernetes cluster. So what should you do, and when? Let's have a closer look at the different options.
+> If this is the first page you visit, have a look at the [introduction page](/) before moving on.
 
-The first thing you need to consider is the environment, in which you want to deploy your Kerberos Agents. Is it inside your house, your company or at a public place? Another important factor is the number of camera streams you would like to monitor, and if you have any high availability requirements.
+## Where to start?
 
-## Kerberos Agents
+Starting something new is not easy, there is always a steep learning curve. While setting up the different Kerberos.io components, you might ask yourself:
 
-The most common scenario is to install your Kerberos Agents at the edge. This has a couple of benefits such as latency, bandwidth and a less complex security model. When installing one or more Kerberos Agents at the edge, you have different options.
+- How should I deploy these so called, [Kerberos Agents](/agent/first-things-first/)?
+- Do I setup a [Kerberos Vault](/vault/first-things-first/) next to my [Kerberos Agents](/agent/first-things-first/) or in a managed cloud?
+- Can I self-host [Kerberos Hub](/hub/first-things-first/) or do I need to install it in managed cloud?
+- And probably many more questions..
 
-### Less than 5 cameras streams
+We will discuss some of the most common setups we have seen, however this doesn't mean that your setup will not work if it's not shown identically in this section.
 
-When having limited number of cameras streams, and there is no need for any high availability or scaling, then multiple individual Kerberos Agents are a good choice. For example, you can install the [Kerberos Agent](/agent/first-things-first) on a Raspberry Pi, a VM or install it from source.
+## Home setup
 
-On below architecture, a couple of Kerberos Agents are installed at the edge (inside the local network). A Kerberos Agent is installed for each camera stream, and each Kerberos Agent runs on a separate host (Raspberry Pi, VM, etc). The front-end of the Kerberos Agents can be reached from the internal network, but cannot be accessed from the internet.
+If you are starting with a small deployment - for example for your home - than probably you want to keep it simple. Either you'll only host a couple of [Kerberos Agents](/agent/first-things-first) or an additional [Kerberos Vault](/vault/first-things-first/).
 
-{{< figure src="kerberos-agent-edge.svg" alt="A Kerberos Agent is monitoring a single camera stream." caption="A Kerberos Agent is monitoring a single camera stream." class="stretch">}}
+{{< figure src="deployment-home-setup.svg" alt="The home-setup deployment" caption="The home-setup deployment" class="stretch">}}
 
-Another option to install the Kerberos Agent is to use Docker. Instead of having separate hosts for each Kerberos Agent, you will have a single host (Docker host) with multiple Docker containers. A Docker container is deployed for each camera stream, which contains the software to run the Kerberos Agent.
+In a home setup you'll probably rely on `docker` instead of `kubernetes`, mainly because of simplicity. In this setup, [Kerberos Agents](/agent/first-things-first) are installed on a compute at the edge, next to a [Kerberos Vault](/vault/first-things-first/). For both components you can leverage `docker` or `docker compose` to setup and configure the system.
 
-Running your Kerberos Agents inside Docker containers, allows you to process multiple surveillance cameras using only one host.
+The advantage with [Kerberos Vault](/vault/first-things-first/) is that even if it's deployed at the edge, you can still target a cloud storage system like S3, GCP Storage, etc. Next to that the main advantage is speed, you typically setup an [Kerberos Agents](/agent/first-things-first) within 5 minutes, and a [Kerberos Vault](/vault/first-things-first/) installation within 30 minutes using `docker compose`.
 
-{{< figure src="kerberos-agent-docker.svg" alt="Having just a few of cameras to be monitored then Kerberos Agent on Docker might be the right choice." caption="Having just a few of cameras to be monitored then Kerberos Agent with Docker might be the right choice." class="stretch">}}
+## Cloud-first setup
 
-### More than 5 camera streams
+One of the most common setups is the **cloud-first** setup, where you install the majority of the components in a managed cloud or your own private cloud.
 
-Having more camera streams and Kerberos Agents will become a bit tedious to manage, and you might start looking for other solutions in terms of scaling and also storage.
+{{< figure src="deployment-cloud-first.svg" alt="The cloud-first deployment" caption="The cloud-first deployment" class="stretch">}}
 
-When having more than 5 [Kerberos Agents](/agent/first-things-first/), the Kerberos Enterprise Suite starts to become interesting. While building on Kubernetes, your Kerberos Agents will be scheduled and deployed more easily through [Kerberos Factory](/factory/first-things-first/). Your recordings will be stored centrally using [Kerberos Vault](/vault/first-things-first/), and you will be able to review everything from within a single pane of glass with [Kerberos Hub](/hub/first-things-first/).
+The huge benefit of this approach is that your [Kerberos Agents](/agent/first-things-first) are installed next to the camera infrastructure, and ideally in the same network. This will bring latency and data transfer to a minimum.
 
-{{< figure src="kerberos-enterprise-suite.svg" alt="When you start having bigger deployments it's recommended to go with a Kubernetes approach." caption="When you start having bigger deployments it's recommended to go with a Kubernetes approach." class="stretch">}}
+The [Kerberos Vault](/vault/first-things-first/) is installed in the cloud together with some scalable cloud storage. The [Kerberos Hub](/hub/first-things-first/) is installed in the same or other cloud as the [Kerberos Vault](/vault/first-things-first/).
+
+The benefit of this deployment is that you only need to install [Kerberos Vault](/vault/first-things-first/) and [Kerberos Hub](/hub/first-things-first/) once. Most of the work is setting up the [Kerberos Agents](/agent/first-things-first).
+
+Another benefit is that is latency and bandwidth consumption, as we have compute at the edge for running the [Kerberos Agents](/agent/first-things-first). Only data (recordings) that are relevant will be send over the network to [Kerberos Vault](/vault/first-things-first/).