Join us for the Parallels RAS Technical Foundations Series, a four-part technical webinar program designed to help IT professionals build the knowledge required to deploy, operate, and scale Parallels RAS with confidence.
In this first session, we’ll start from the ground up.
Before launching the installer, it’s critical to understand the architecture you’re building. The decisions made at the design stage determine how well your environment will scale, perform, and handle future growth.
This session walks through the core architecture of Parallels RAS, explaining how each component works and how they fit together to deliver secure, scalable virtual application and desktop environments.
Whether you're preparing for your first deployment or reviewing an existing design, this session will give you the technical foundation needed to move forward with confidence.
In this on-demand webinar, we cover:
- Core Parallels RAS architecture and design principles - Key components of Parallels RAS, including Secure Gateway, Publishing Agent, Connection Broker, HALB, and Session Hosts - Deployment prerequisites and system requirements - Licensing models and how they impact design decisions - Best practices for planning a scalable, production-ready environment - Common deployment mistakes and how to avoid them
This session is ideal for:
- System administrators deploying Parallels RAS - IT architects designing application delivery environments - Partners implementing Parallels solutions - IT teams preparing for Parallels RAS Technical Professional certifications
AI Generated·May contain errors
Video Summary
AI Generated
This technical webinar series provides a comprehensive overview of Parallels RAS, a turnkey solution for delivering secure remote applications and desktops. It covers the system's architectural components, deployment flexibility across hybrid clouds, multi-tenancy capabilities, and a simplified single-SKU licensing model designed for scalability and ease of management.
Welcome to this first out of a series of four technical webinars where we're going to cover
all the content of the Parallels RAS technical professional training. My name is Steven de
Winter. I'm the Global Senior Director for Sales Engineering and we have a packed agenda in this
first series. So what we're going to cover is everything you need to know, let's say from
a theoretical point of view. So we're going to talk about what Parallels RAS is, we're going
to go over all the components, we're going to look at the requirements and then different
architectures, talk about licensing and then I'm also going to point you into the direction of all
the technical resources and then in the next webinar we're going to do like the real stuff,
start the full installation from scratch. But for this webinar it's all about everything you
need to know before you start. Okay, so for those of you who are really new to Parallels RAS,
let me start with giving you a quick overview and a quick demo and then we're going into more
the technical parts. So Parallels RAS is a turnkey solution, it's a solution that contains everything
you need to offer remote applications and desktops to any user and any device in a secure and simple
and easy way. From a user point of view, and we're going into more details later on, we have
several ways that we can connect to the system and we can do that fully browser-based but there
are also like native clients for Windows, for Mac, Linux, Android, iOS and then there's also
the possibility for example to integrate with different clients from different vendors.
On the resource side, we're covering also all kinds of use cases, so we can
for example offer desktops and applications from Windows session hosts, typically a Windows server
where we can connect to and be connected with multiple users at the same time. But you can do
the same thing with individual Windows 10 or 11 desktops, we can take over even physical machines,
integrate it as a virtual desktop and also work with local applications. Also on the resource
site or the use case site, we have a lot of use cases that we cover and then the Parallels RAS
software itself also contains several features, functionalities and components and this is what
we're going to discuss more today, is all these layers. So we have a built-in user portal,
load balancing capabilities, single and multi-tenant. We can integrate into different
hypervisors and clouds on-premise and cloud, single sign-on capabilities, auditing, monitoring,
reporting and management. So we're going in this series of four to go over all these parts and
explain how they work. Before I start with explaining the details, let me do the two,
three, four minute quick demo of Parallels RAS for those of you who have not seen it
before and then once you have seen it, we'll start explaining all the parts and how it works
and what is happening here. Let me just start with giving a quick demo of the web client.
So for example, my current environment that I'm using is called connectparallelsras.net.
I can go with any browser to this environment. I can log in with my username and password.
In this case, my username and password is my Windows Active Directory username and password.
In general, we would say that Parallels RAS requires Active Directory to run fully functional.
It is also technically possible to operate it in a workgroup, but then it has, of course,
some constraints. Once you are logged into the system, what you will see is that you get access
to a list of applications and desktops in my case. So this is something you can mix and match
based on different profiles or different conditions. This is something we're going
to cover in one of the next webinars, how to configure everything in detail.
But in the most simple case, what you can do is you can go into the system and then
there are two ways, for example, we can start this application. I can either
open it completely in the web browser or I can open it through my Parallels RAS client.
I'm doing this series from my Windows laptop. So let's just use, in this case,
the Parallels client for Windows. So if I say open Parallels client, what you will see is that
it will open my client and then the application should start in a few seconds. So that is the
banner already. The session has been started and then you give it a little bit of time. There it
is. So the Microsoft Report Builder is there. So as you can see, very simple to bring an application
to any device. What we also can do in here is then, for example, open files on my computer.
We can also do things like printing. There are features and functionality to redirect webcams or
other USB devices if that would be needed. Also something I'm going to show you right now,
but we're going into much more detail later on during this series, is the Parallels Management
Console. So as an administrator, I also have a tool called the Parallels Management Console.
And in that Parallels Management Console, I can do all the configuration. For today,
we're not going to spend a lot of time into this tool. But for the next webinars,
we're going to spend a lot of time showing you how everything works in here. But this is
basically it. So as a user, you can connect either with the web client or with your native
client. So this is, for example, the native client. As you can see, I have the same applications,
I have the same desktops, which are available to me. And then from here, I can run them.
Let's switch back to the slide deck and bring it back onto the screen. There it is. So
after this very short demo of how it works from a user point of view, let's switch to the
admin side. And let's go into a little bit more detail about all the features,
the functionality, and then all the components to get into an environment similar like the one I
just demonstrated. So the first thing I want to say is that Parallels ROS is a very feature-rich
product. And it's all built around three values that we keep very high at Parallels.
It's all built around simplicity, flexibility, and security. And that, of course, for an
interesting total cost of ownership and, of course, taking user experience into account,
because as with every product, if the user is not happy, then it's not a good product.
And I'm not going over all these items one by one. Some of the things I just want to highlight
are, for example, everything we're going to do in the next series of webinars can be done
through PowerShell and through API. We have also a webinar already done on how the API
and PowerShell work. So if you're interested in that part, please have a look at that recording
from the previous webinar. Another important topic is, for example, the product can be
configured also in a single and multi-tenancy way. For this TP series, so for this series of
four webinars, we're going to focus on a standard installation, so on an installation where we
cover everything from a single tenant. But there is also already webinars available going into
detail about this multi-tenancy and how to configure multi-tenancy. We also have deep
integration into products like MSLogix and MSIX AppAttach. That's something we're going to cover
in the next webinar when we start publishing applications.
On the flexible side, I already mentioned it on the previous slide,
we cover all kinds of cases going from managed devices to unmanaged devices,
cases from offering applications to full desktops, cases where we integrate in cloud,
where we integrate on-premise or where we integrate hybrid. So we offer you the full
flexibility to integrate and to work on any device possible. From a cost point of view,
this is something we're covering in this webinar. I'm going to talk about licensing,
but it may be already interesting that you know that we have like one single license,
one single SKU, and everything is included. So this also makes it more easy when I'm going
later on over the architecture and all the options and the components. Just know that
all the components, all the options, all the possibilities are part of the standard license.
User experience, you have seen it in the two, three minute demo that everything is there to
make sure that users can work efficient and easy. For example, if you're using like our clients,
we have capabilities like drag and drop or printing and scanning capabilities. Everything
is there. We're going into also the next webinar. And then security, that's something we will cover
mainly in the last and the fourth of this series of webinars. But the product comes with a whole
bunch of built-in capabilities and features. So we can, for example, integrate via single sign-on
into external IDPs. But there's also capabilities for built-in MFA, for example, to a QR code that
you scan with your Google Authenticator or Microsoft Authenticator. We have done integration
so that these are all things I'm going to demonstrate in the future webinars of this series.
But let's continue. Let's have a look at all the components and everything you need to know before
we can start the deployment. And I'm going to start with explaining how the overall architecture
looks like and how things are connected to each other. So let's start with saying that there are
two main components. I'm going into more details on these later on. But the two main components
that are required in every installation are the secure gateways on one hand and then the
connection broker on the other hand. The secure gateway is, let's say, the entry is where you
connect with your either web client or with your native client. And that's like your entry door
into the system. And then the connection broker, that's the brains of the system. So the connection
broker is, for example, the system that will do authentication or that will, for example, say like
Steven wants to start this application, please connect to that session host. You can start it.
And to make the gateways high available, there is also a component called the high available
load balancer. We're going also into more details on that later on. But that's something that comes
with the product. And if you want to use a third party load balancer in front of your gateway,
that's technically possible. But also know that there's a built-in solution that can be used
high available. And then once you are through the secure gateway and you're authenticated and
you know what needs to be started, there are several options. There is a single session
host that can be started like a Windows 10, Windows 11 for desktop. We have capabilities
on multi-session and then on the Windows server, you can, for example, start a remote application
as I just demonstrated. And then lastly, we can also take over things like a physical PC. So if
you have somewhere a machine on a desk that needs to be taken over remotely, that's also
integrated. And one of the pieces which is also super important, and that's something we're going
to cover in detail in the third webinar, that is what we call the providers. So the providers
are the integration points we have into different hypervisors or clouds. We have a first series of
seven tier one providers that integrate into the major players in the industry. So we have
integration into things like VMware, Hyper-V, Nutanix, Scale Computer, to on-premise
integrations. And then we have also integration capabilities into Azure,
also AVD, and into Amazon EC2. Something which is new since RAS 21.2 is that we have now also
a custom provider frame. That's a sort of provider that will allow you to integrate via scripts and
other integration points into any cloud, any hypervisor, or any API. We're also going to
cover that later on, but that's also something which is new in RAS 21.2. And of course,
one site is nice, but many organizations or many partners require multiple sites. Maybe you have
like a site in Azure and a site on-premise, or maybe you have a site on the east coast of the
US and another site on the west coast. So all should know that Parallels RAS supports multiple
sites and that all these sites can be managed from a single management. Let's go a little bit
more into details now, and let's switch from, let's say, a logical architecture into a more
technical architecture. So if you take a look at the different components, we need to discuss a few
things. We need to discuss concepts like farm sites, brokers, gateways, and so on and so forth.
So let's go one by one over all these components and let me explain to you what you need to know
about these components, what the roles are, what the options are. And then once we have that,
then we're going to start making a puzzle and starting to make different architectures and see
what can be tested or what can be done with each of these architectures.
The first concept that I would like to explain is what we call a farm. So a farm is a logical
structure that manages all RAS components. So a farm is, let's say, the biggest
unit of resources. Whenever you create the farm, you automatically will also create the primary
site. So whenever you deploy your farm, there will always be at least one site. You cannot have
a site without a farm, but you can also not have a farm without a site. So there's always at least
one site in a farm. The farm is created automatically at the original installation.
So in the second webinar, we're going to do an installation. The installation starts by
installing the primary connection broker on your first machine. As soon as that first connection
broker is installed, it will automatically create the farm and the site. Important to know is that
you need a license per farm. So if you install a license, the license will be on the farm level,
not on the site level. So that also, for example, means, and we're going into more detail later on,
the licensing is, for example, based on concurrent users. If the same user logs
into site one and he logs at the same time into site two, still that will be counted as one
license. The initial user that has done the installation of the primary connection broker
in the first site in the farm will also be called the root administrator, and that's the user that
has two admin rights over the overall system. We already discussed sites briefly, but within
a farm you have at least one site. You can create multiple sites, for example, to isolate resources
or to delegate permissions. One of the reasons that you want to have multiple sites could be
maybe you have a data center A and a data center B, or you're a service provider and you want to
create different sites per customer. It's possible to create as many sites as you want.
Your first site will also be called your licensing site. So it's the first site that, for example,
will take care of the connection with the license server for validating if the license is still
valid, or, for example, when we're using service provider licenses to communicate and report on the
usage. There are some features in which you can replicate certain objects between sites. So,
for example, what you could do is maybe two sites, a data center A and a data center B.
You could, for example, configure that whenever you add something to the published applications
on site one, that same resource is automatically published on site two. We cannot replicate
everything. So, for example, in the advanced architecture webinar, we go into more details,
but, for example, the session hosts or the host pools, which we're going to explain later on,
are always part of one site. So if you, for example,
set the replication on and you publish, for example, an application, that application will
also be published in your second site, but then you still have to link it to the corresponding
pool on the second site to finalize the application. The most important component,
as already mentioned, are the connection brokers and the gateways. So the main role of the
connection broker is to, first of all, take care of the authentication. It also takes care
of the load balancing of the published applications and desktops. So, for example,
if you have multiple session hosts and you want to start an application, it's the connection
broker that will, for example, determine which of the machines is not heavily loaded, and it's
that system that then will decide, for example, to go to session host number two or session host
number five to start the applications. And it's also the central source of all communications with
other services, and it also manages the system settings. So this is maybe something important
for administrators who are used, for example, to deploy systems like from Citrix. There is no
centralized database, so you don't need a Microsoft SQL server for the configuration.
The configuration is stored on the connection broker, and then when you have multiple connection
brokers, the configuration is automatically replicated between the different connection brokers
whenever you make a change and you apply that change. Also, already as mentioned, you start
the installation by deploying your first connection broker, and then from there,
the rest can be automated. We're going to explain that during the next webinar how that
all technically works. For redundancy reasons or for scalability reasons, we can, of course, also
have multiple connection brokers per deployment. So what we could, for example, do is we could,
for example, have one primary connection broker and then a secondary connection broker for
high availability. Linked to the licensing is the primary connection broker that, for example,
will take care of the licensing. It's also that primary connection broker where the
changes are applied first, and then from there, they automatically replicate to the other systems.
Secondary connection brokers are there for availability, for scalability, and when needed,
they can be promoted either manually or automatically to become the primary connection,
because it's important that your primary connection broker is back within 72 hours. If it
could not be back in 72 hours, the license could not be validated, and then the system
would stop working. So there are possibilities, for example, for an automatic auto-promotion
of your secondary connection broker if needed. And as mentioned before, the configuration is
stored on the connection broker and then replicated amongst the different connection
brokers automatically. And this is also to avoid, for example, that you have to install and configure
expensive SQL servers or, for example, to actually think about high availability,
everything is done automatically. Also, in general, and that's mainly a recommendation for,
let's say, for the more large environment, so if you go above a certain amount of
users, let's say you go above 1,000 users, then we also would recommend to do like a sort of
n-plus-one setup. If you have two brokers for availability, install a third one,
so that whatever happens, there is always one spare broker that can be promoted or that can
be used to make sure that everything is working. Second component are the secure gateways. So the
secure gateways are the entry points for the RAS farm or site. It would typically tunnel all the
RAS traffic through a single port. Although I have to be careful, you can also configure your
client in direct mode. So if the client, and we're going to discuss that also later on, if the client
is, for example, part of the same network as the resources, you can configure the client to go into
direct mode. And in that case, the client will first connect to the gateway for authentication,
but then once the authentication is done, the traffic will go directly to the resources. So
it can tunnel all the RAS traffic through a single port, which would be typically the case
if you're connecting, for example, from public internet to resources inside your network, but
it's also possible to have other deployment scenarios when needed. It can also take care of
SSL encryption and decryption. And it's also the place where the web client runs.
So the web client runs on the secure gateway. Web client is an optional component. So the
web client is not mandatory, but if you activate the web client, it will run on a gateway.
The gateway can be configured in two ways. Normal mode means that it will just take a connection,
handle it, talk to the connection broker and do everything which is needed. It's also possible to
configure a secure gateway in what we call forward mode. So that means that it will arrive on the
gateway and then it will forward to another gateway. So this is something I'll reveal,
for example, if you want to have like a gateway and a DMZ network, and then, for example,
a second gateway inside your internal network. And then from there, we can configure everything
that traffic first goes to the gateway and the DMZ, and then it's forwarded to a normal
gateway in the internal network that then speaks, for example, to the connection broker.
These are typically advanced architectures that we already have discussed in other
webinars. So we're not going into details during this basic DP setup for webinars,
but if you want to know more about that part, that can also be looked at in that webinar.
Also, similar like the connection brokers, we recommend to have like multiple gateways for
high availability and redundancy and scalability. And then the web portal, as mentioned before,
it's optional. The main advantage of using the web portal is that any user can connect without
having to install any clients on his device. It's HTML5 and it will allow you to connect
natively with your vendors. On the resource side, everything works with agents. Parallels
has multiple agents that can be deployed. If you deploy on a Windows server operating system,
then typically you deploy the RDS session host agent. That's the agent that you use for
multi-user Windows sessions. That also requires Microsoft RDS licenses. I'm going to cover that
also a little bit later. But we can simplify a lot of things on the parallel side, but unfortunately
not the Microsoft licenses. So if you connect to a Windows server, you need Microsoft RDS
licensing. So if you're using the RDS session host, that is where they are required. It can
be used both for individual applications or for desktops. And it allows you to do things like
auto-scaling. Auto-scaling is something we will cover in the third webinar when we are discussing
these templates, clones, and providers. The VDI agent is typically for Windows 10 and 11. We also
have a VDI agent for Azure Virtual Desktop. These virtual desktops can be persistent and
non-persistent. So we can, for example, say whenever Steve logs in, always send him to
the same desktop. Or whenever Steve logs in, we send him to a boot desktop. So that's one option.
And then we can make them persistent or non-persistent. That means that
whenever I log in, I have an empty desktop. Because whenever I log out, this whole system
is reset back to the original state. So all options are available. Remote PC doesn't require any
cloud or hypervisor infrastructure. It's just connecting with the machines that are there.
There is also features and functionality for things like break on launch. And then Azure
Virtual Desktop, that's a special agent that we have developed to integrate directly into Azure
Virtual Desktop. So we can also do things like provisioning of all the components and everything
on the Azure Virtual Desktop side. And it also does things like full cost optimization. So,
for example, we can, when you shut down an Azure Virtual Desktop host because it's not needed,
because the load only requires one machine and you might have maybe four machines,
we can move that machine to low cost storage and then bring it back to expensive storage.
Another concept I have to cover, and that's going to become important when we are discussing and
showing how to publish applications. Typically, you publish an application not from a session
host, but you publish it from a host pool. So a host pool is a group of session hosts or VDIs.
It's a group of resources. Important to mention is that the resource can only belong to one pool.
So if you have a VDI or you have a session host, it can be only part in one pool.
There are two types of pools. We have what we call static pools. So static pools are pools
from existing machines. You create them manually and then you assign them to a static host pool.
The other way it can work is based on templates. That's something we will cover in the third
webinar when we're discussing the providers that we can. For example, say like this is the template,
create now 10 clones from this template and put these 10 clones in a pool and whenever more
resources are needed, add machines whenever the resources are not needed anymore. The main
functionality of the host pool is load balancing. It's to make sure that if your users connect that
all the resources are distributed equally over all the users. Other things we can do in host
pools are things like updates, power management, redeployment. So it's also a level of resource
management and power management that can be done. We'll continue. So providers we already touched
and we're going into way more details during the third webinar. The main role of the providers is
to integrate into the underlying systems, taking care of things like life cycle management,
start, stop, reboot, clone, scale up, scale down. And it's essential, for example,
in the context of VDIs or in templates for auto-scaling. So this is the native thing
that makes sure that we can do all these automation and integrations into the underlying
platforms and systems. And there are two ways these providers can operate and work.
Either the providers or the provider agents will run on the connection broker. That's the
default behavior. So if you, for example, add a provider for VMware or for Azure, it will typically
run on the connection broker. Another way it can work is from a dedicated host. So
we can also run the provider from a dedicated host and the connection broker will speak with
a dedicated host and that dedicated host will have the provider agent that will speak, for example.
Why would you like to do that? Maybe you want to have your VMware console only accessible
from a management plan and you put your connection broker in another network, then
you only have to open one firewall port into your connection broker and this
provider agent and then a provider agent can help you de-center in that dedicated management.
And when it comes to the integration, and this is something new also from
ROS21.2, we have two levels of integrations into
Clouds and Hypervisors. So the tier one integration, or what we call the native
integration. So this is a list of seven systems that we support going over VMware, Hyper-V,
Microsoft Azure, Amazon, also AVD. And then this is also something new, so we already support
Nutanix AHV for a long time, but also since ROS21.2 we support Prism Central,
scale computing and then of course the remote PC provider. And then also new in ROS21.2,
we have this open provider framework that we can then, based on some scripts and other things,
integrate it to any Hypervisor or Cloud. So think about integration into Proxmox,
with the rules of OpenStack, whatever Hypervisor you might have or know technically possible.
Same thing with whatever Cloud, we have native integration to Amazon and Azure, but maybe you're
a Google user or an Alibaba user or an Oracle Cloud user, also with this open framework we
allow you to do automation with these platforms. Same thing with an API, maybe you're a large
managed service provider with your own interface, we can also, for example,
integrate that API and, for example, let your system operate from in-parallel as console
and create and flow machines. Next part, components, we're almost through the list
of components and we can have a look at the architectures. So the next component is the
High Available Load Balancer or HAL-B. Until now, all the components that were discussed
were components that you had to install on Windows Server infrastructure. We're going
over these details in the next slide. But the HAL-B, it's a Linux Virtual Appliance, it's a VM
that you need to deploy on your infrastructure. And its sole purpose is to load balance the traffic
over multiple gateways. How this works and how to set up is something we will cover in the fourth
webinar of this series. But it's very simple, you have a virtual IP address that can, for example,
flip between two of these HAL-Bs and then the load is distributed over the different gateways.
From a management point of view, I've already opened it, but we have on one hand the Parallels
RAS console, that's the console where you can do everything and manage all the systems as,
like I say, a full administrator. There is also a web management portal. So if you want to
delegate certain rights to a HAL desk or you want to do some lightweight actions, then this is
something that can be done through the web portal, which is also available to admins.
Clients. So you already have seen the web client, you already have seen the Windows client,
there is also a native client for Mac. We have dedicated documentation on all the clients,
what works, which features, what functionality are supported. Also iOS client, Android client,
something which doesn't exist anymore is the Chrome client. So we used to have a client
for Chrome OS that's not available anymore, that has been deprecated. If you're using Chrome OS,
we have a very powerful web client which is available for use. On the client side,
requirements, and I think it's important from time to time to look at the documentation. At the time
of this webinar, following clients were supported. So Windows 10 11, Windows Server from 2012 all
the way up to 2025. Mac OS 13 till 2026. We both support the clients running on Intel as on the
native Apple silicon processor, so that's possible. For Linux, we support different
distributions. Have a look in the documentation to find out exactly which version of Ubuntu,
Debian, Fedora and Linux we support. Same for Agile, both support for Agile OS 11,
Agile 12 here in the marketplace. We are also compatible with other 10 client providers,
like for example, Tenzic and a few others. So have a look at the documentation to figure out
which clients are supported. And then of course, we also support recent versions of iOS, iPadOS,
and Android. And to repeat again, we used to have a client for Chrome OS,
that one has been deprecated. We recommend to use the web client as for now.
Let's switch from clients and let's go to the backend. Let's look at the hardware requirements.
Of course, the hardware requirements depend heavily on the amount of users and the type
of applications and the type of use cases you want to support. As a minimum, either physical
or virtual, we would say two CPUs, four gigabytes of memory. If it's a small test,
a small deployment or a test deployment, let's say lower than 30 concurrent users,
in theory, you can run everything on a single host. That's up to you. Of course,
you will not have high availability at that point, but it's technically possible to put
the session host and the broker, for example, on a single host for less than 30 users.
If high availability is important, or you have like a larger setup, then
we would heavily recommend you to deploy multiple Windows systems and then of course,
distribute the components of these multiple systems. Options on how to do that, we will
also discuss later on in this webinar. The sizing of the session host of the systems that
run the applications and desktops is, of course, much more depending on the amount and the type of
applications and the kind of user workloads. This is a Microsoft recommendation. This is
not a Parallels recommendation, but of course, we take it over. This is based on what Microsoft
recommends. For example, if you're having like a light workload, then you can have, for example,
six users on two CPUs, 80 gigabytes of memory. Of course, if you're like a power user and you're
doing like 3D graphic designs and you have things like AutoCAD open or things that require CPU and
memory power, then of course, we would start recommending having much more memory and CPU.
If you take a look on the OS requirements side, I didn't put all the components in this list,
but it is in our documentation. You can find all the details. I'll show you at the end of
the webinar today where you can find all the information. But in general, let's say all the
core components like brokers and gateways, we expect those to be at least server 26 up to 25
and both the core or the desktop experience can be used. For the provider agents,
same requirements. When it comes to session hosts, so if you're using a Windows Server operating
system as a session host, you can go lower, so maybe support up to server 10, 12, R2 and upwards.
If you're having 26 or newer, you need to install the desktop experience options. If you would be
using 2012 R2, then please make sure that you don't install the server core version because
that will only support it for the session host deployment. It's supported for a connection
broker or for a CPU gateway, but not for a session host. And then the VDI agent or the guest agent,
that works of course on Windows 10, Windows 11. You can also install it on a server operating
system if you want to connect with a single user. Now that we have all the components,
let's go a little bit deeper into how to design and architect an architecture.
So there are, I think, three key considerations. Of course, the amount of concurrent users is
important. It's different from designing an architecture for 30 users versus, for example,
3,000 users. Same thing with high availability. Maybe you only have 30 users, but these users
need to be always on. Maybe it's healthcare or another important use case. Then, of course,
you need to take a look at redundancy, high availability. So that's a second consideration.
And then the last one is network security. Maybe you have strict rules that says users need to be
in my DMZ before they enter, or maybe it's a use case where you're just offering it from
an internal network. Nothing is exposed on the outside world, and it's not problematic.
So network security could be a third consideration on how to architect and design your
infrastructure. These are some best practices. Let's say these are not like hard limits,
but things that we would recommend to take into consideration.
We would recommend not to go above 5,000 concurrent users per site. If you have a very
large environment, we would recommend, for example, to put 4,000 users in one site with
the other 4,000 users in the second site. Also, for larger sites, we would recommend to go for
three connection brokers. On the gateway side, 500 users per gateway is like a sort of
recommendation. You can go higher. You can go to 1,000. I would say 1,200 is like a sort of
maximum. Same thing with the highly available load balancer appliance. We would say go to 2,000
concurrent sessions, not more. And then, of course, you need to add more session hosts or VDI
hosts to scale the workloads and the usage. Let's have a look at some architecture
possibilities and designs. Let's start with the most simple one. Just throw everything on a single
box. The connection broker, the secure gateway, and the RDSH agent is on the same box. This will
allow you to test applications, test compatibility of devices, test the clients, test printing,
client policies, and all these kinds of things, MFA, authentication, single sign-on. You can test
all of these components in this setup. Of course, if you only have like one box,
things like load balancing won't work. Network segregation won't work. Redundancy won't work.
If these things are important, then what you will have to do is at least start decoupling
things from each other. For example, the first thing you could do is keep the connection broker
and the secure gateway on one machine and then put your applications on a second machine. This
will, for example, help you to scale. You, for example, you can now have like maybe 100 users
using that same application because you have like maybe like two or three session hosts to
distribute the load on the backend. And it will also allow you, for example, to
segregate certain roles from each other. So, the gateways and brokers on one side and the
session hosts on the second side. The first level of introducing high availability would be, for
example, to deploy two boxes, install the gateway and the connection broker on each of them. So,
you have a primary connection broker and a gateway on the first box, the secondary connection broker
on the secondary box. You could then, for example, deploy the hull base. Of course,
if you only deploy one, you create a new single point of failure. So, we would then recommend to
install two. And there is like a virtual IP address that will load balance the incoming traffic
amongst these high available load balancers. And then these would then distribute to the gateway.
So, the hull base would be active-passive and the gateways would be active-active. And then
session hosts. If you want to split up more and network segregation is super important for you,
you could, for example, say I put my gateway on one box, I put my connection broker on another box,
and then I have my session hosts on a third set of boxes. So, that's another way how you can split
things up. And then, of course, if you combine number three and four, you could also make,
of course, things high available. So, let's say that this architecture would be a sort of,
let's say, reference architecture for an environment with a few hundred concurrent
users where you want to have everything isolated from each other and everything
high available. And you would say, like, deploy to hull base, to gateway, to connection brokers,
and virtual session hosts. And then from there, you can start scaling things up around the need
for more users, more connections. But this would be like a simplified architecture taking
into account high availability and network segregation. Good. These are basic architectures.
We already have done a webinar on advanced architecture. So, for example, if you want
to know more about these double-hop gateways where one gateway sends things to another gateway,
that's something we have covered in this previous webinar on advanced architectures.
Same thing with multi-site deployments. So, we have no state within one site. But if you want
to deploy, like, a second site, then you would need to have, like, another pair of connection
brokers and another pair of session hosts. Important to mention is that resources can
only belong to one site. So, whatever you deploy, a host, a gateway, a broker, they are always part
of a single site. And then the multi-tenancy, that's a setup where you have a tenant broker
and then multiple environments behind. You could, for example, be a managed service provider. You
have, like, two customers there. You have a customer Coca-Cola. On Azure, you have a customer
Pepsi-Cola in their own data center. There is a possibility with the tenant broker to say,
like, if somebody from Coca-Cola connects, send them to the Coca-Cola setup. If somebody from
Pepsi-Cola connects, send them to the other setup from Pepsi-Cola. So, multi-tenancy is also
something we cover in this advanced webinar that you can already look at, which is also available.
Good. With these components covered, let me switch to another important topic before
we can start the deployment in the next webinar. That's licensing. As mentioned before,
licensing is very simple and straightforward. So, it's one license, all features included,
and it's based on concurrent users. So, if you want to deploy per user and separate
each with two gateways and two connection brokers, that's fine. That will cost you a lot of
resources and a lot of hardware, but that has nothing of an influence on the Parallels RAS
license. It's purely based on the amount of concurrent users. So, there are no hidden modules
or feature upgrades. There are two models. So, we have the classic subscription license.
This is typically where if you're like a partner and you're into reselling and the end customer
becomes like the owner of the license, then it's typically like a subscription license that you
sell. These can be purchased in a one, two, three, up to five-year subscription model.
Important is that after one to three years, the license will end and you have to renew it.
There is a 30-day grace period, just to be aware of that. The second option is
something specific for partners. We have a service provider license agreement
where it's subscription licenses. If you, for example, would go for a three-year subscription,
that would be like three years that you pay upfront. With the service provider license
agreement with Splunk, it's cost paid. So, what we will do is we will take a look at the
consumption of the end of the month and we will invoice you for the used licenses. Again,
it's concurrent user-based. It also works with tiered pricing. So, if the price for 100 users
will be per user higher than, for example, if you reach 1,000 users or if you reach, for example,
5,000 users, the price will decrease based on the amount of concurrent users that are in that
month. There is also a possibility for pre-commitments. You could, for example,
say I want to pre-pay 500 licenses and then as long as you're not going above 500 licenses,
it will be covered with your pre-commitment. The moment you go above,
then you will need to pay on the delta on that monthly basis.
If, for example, you would use sub-licenses in the context of Splunk, that would be all added
together. The same thing with subscription licenses. You can also create sub-licenses there
if that would be needed for some reason. Another thing I want to highlight is that
how are licenses counted? If you connect to an application, to a public application, independent
if you're an administrator or a non-administrator, it will always consume a RAS license. If you
connect with the RDP client or with a RAS client or with a web client and you're an administrator
or a normal user, you will always consume a license. The only way you will not consume a
license is that if you connect as an administrator to the system, either via RDP or via the RAS
clients, at that moment we will not account the usage as a licensed usage. Also linked to the
licenses. License management, and we're going to cover that in the next webinar, is done through
the MyParallels system. The MyParallels system is where you can have access to your licenses
and if you want things, I'll cover that on the next slide. Internet connectivity is required
for ongoing validation. There is a 72 grace period time, so if for some reason your network
will be interrupted or your primary connection will be down, as long as you get back in three
days, it should be no issue. There is also capabilities for overusage. That's maybe also
something interesting to know. Even if you have a subscription license of let's say 500 users
and you're growing and you didn't notice that you're hitting the limit of your license,
for seven days you can go over. You can go from 500 to 600, 700, 800, whatever your new number
becomes, but then within seven days you have to resolve the overusage by either
decreasing or by buying more licenses. Also important to mention is that
licenses, except that Parallels is a single license, everything is included.
Of course, that excludes some third-party licenses, like I already mentioned the Microsoft
SQL, but also the serverless calls or SQL server licenses, so I have to put a little nuance with
that one. I've mentioned that for the installation and the configuration of Parallels, you don't need
a SQL server. There is a component which is called our reporting services, so all the telemetry,
all the usage, all the details can be stored in a SQL server reporting server. That's something
we will do and cover in the fourth webinar. That requires SQL server, but that can also be done
with SQL Server Express, so it's not mandatory to have a paying SQL server, but if you do,
of course, you need to take it into account. Same thing like for Hypervisor licenses or
Cloud Platform licenses, these are also licenses. This brings me to the last topic to
discuss today, and then I'm going to have a look at the questions that came in.
For those of you who want to read more or find out where to find everything, because the next
webinar is going to start with deployments, there are three important pages from a technical
point of view. First of all, docs.parallels.com, so that's publicly available, no passwords,
no accounts, nothing is needed. You can just go there and have access to all the
manuals, all the documentation. There is the admin guide, but there's also guides on API,
on reference architectures, on benchmarks that have been done on different sizes, so lots of
very interesting documentation there. For our partners, we have our partner portal.
If you look to become technically certified, that's something you can do on the partner portal.
Then the third important part, that's the myparallels.com account or website. That's
where you can manage your licenses, where you can create support tickets, where you can do
product downloads. In the next webinar, that's basically where we're going to start. We're going
to start there, download the software, and then start the implementation. With that covered,
I'm through the content of this first webinar. For part two, we're going to do like a real
installation. We're going to install our first connection broker, set up the gateways. We're
going to add hosts. We're going to make everything redundant. We're going to publish applications
and desktops, and then we're going to show you how we integrate into MSIX. If you want to know
more about these things, please have a look or follow the second webinar. With that said,
let me switch to the questions that came in. I've seen that there was a lot of people in today's
webinar. There are a bunch of questions coming in. Let me try to cover as many as possible
still in the next few minutes. If your question would not be covered live,
my colleague Paul is probably already answering some of them in the chat as we speak.
Let me first take one. Is the seven days grace period a one-shot, or does it reset
once you're below the limit again? It's a one-shot, so it's not that it will automatically
reset. It's a one-shot opportunity that we give to go over.
There are lots of questions coming in. I'll try to take as many as possible.
How would you notice a server needs more resources if too many users are different
programs? This is something we will also show you in one of the next webinars. We have monitoring
services, and we also have something called notifications. We could, for example,
create a notification that if the load of a system goes above a certain amount, that either
we deploy a script or we send an email to an admin for notifications.
There's also a question on the multi-tenancy. Does that create a single web instance with
roads to users to these respective farms? Yes, it's correct. You have one entry port.
Actually, what we do is we create one gateway or one set of gateways that can be used
in front of different farms. If you want to know more about that, I would strongly recommend you
to watch the webinar, which is called Advanced Architecture to be Covered, the multi-tenancy
in that webinar. I'm trying to take some more. There are some people that are asking where you
can find the previous webinars. They are on our partner portal, and they are also on our Parallels
YouTube channel. If you go to YouTube and you look for Parallels, you will find these webinars.
We also have done webinars on API so far, on single sign-on, and a few other things. It's
a bit funny that we have started with the advanced ones and we are now doing the basic series.
Let me see if I can find other ones. There are questions, for example, on can we integrate
physical PCs if they are not part of the domain? Yes, it's doable. You will need to do a few things,
but it's technically possible. I'm trying to not forget anyone. I see Paul already has also done
a lot of them. Let me just take the last one. There is also a question on Linux. The question
is do you also support Linux desktops? At this point in time, it's not on our roadmap. However,
we have, for example, customers that have a mix of Windows desktops and Linux desktops,
and then, for example, publish a VNC player on a Windows box and then use that as a sort of
jump post to go to the Linux. There are, let's say, workarounds, but we don't have at this point
in time any plans to add Linux desktops to the environment. Let me stop for here.
If there are any other questions, I will still try to answer them on the chat. I want to thank
you for following and see you in the next webinar. Thank you.