Skip to content

Cloud AWS Install

Big Picture

These instructions are for setting up Row64 Server, Clients, and Studio in a virtual private cloud environment.

The majority of the content covers the setup of the environment. Once that is complete, the installation process is similar to the local installation and setup.

These are specific instructions for AWS, but a similar process can be followed for Azure or GCP. (The biggest difference will be the compatibility of OS versions and graphics drivers, as each cloud provider has customer drivers from AMD & Nvidia for their environments).

We can provide guidance or check with your cloud provider

#1 Environment Setup

You need to set up and validate your cloud environment, and access your Cloud Console with permissions to edit the various resources (or your IT dept will need to do so):

  • Virtual Private Cloud (VPC)
  • Security groups
  • IAM Profiles
  • AMIs & Custom Instances (pre-configured instances or custom)
  • RDP (for desktops), SSH (for servers)
  • Graphics driver Installation (if your running studio, want gpu accelerated clients, or want GPU compute on the server)

#2 Environment Setup

Access your AWS VPC Console:
https://us-east-1.console.aws.amazon.com/vpcconsole/home?region=us-east-1#vpcs:

Skip this step if your IT department has configured your VPC. Otherwise, click Create VPC at the top right.

VPC Setup

Unless your organization has a specific setup, use the defaults, then click Create VPC at the bottom right. AWS has all the details on options in their documentation.

#2 Attach Subnet

You need to create and attach at least one subnet to the VPC if none exist. Go to the subnet console and click Create Subnet if none exist:
https://us-east-1.console.aws.amazon.com/vpcconsole/home?region=us-east-1#subnets:

Associate your subnet to the VPC from the previous step and use the defaults.

Note:

  • IPv4 address only

  • Only 1 subnet, so IPv4 CIDR for the subnet can be the same as the VPC CIDR that has one block of address for the whole subnet

Click Create Subnet.

The subnet should resemble the following image, with the Auto-assign public IPv4 option set to yes. It will automatically create routing tables and a Network ACL.

#3 Attach Security Group

You will need to attach at least one security group, or create one and attach it to the VPC if none exist. If needed, go to the security group console and click Create security group at the top right.


Name the security group and attach it the VPC created in the prior steps.

You need to add rules for the inbound and outbound traffic, then select the Create security group button in the bottom right.

See the next section for inbound and outbound rules.

Inbound Rules

Add the following inbound rules:

Destination Ports:
Row64Server: 9002
HTTP: 80
HTTPS: 443
SSH : 22
RDP: 3389 & 8443

(The source is 0.0.0.0/0, which allows all traffic)

Outbound Rules

For outbound rules, leave them to the default to all traffic all ports

#4 Create Internet Gateway

This VPC has private IP address that can be used to access machines on the same subnet.

If you want get out to the public internet, create an internet gateway with the top right button, if one does not already exist.

This example uses the defaults:
https://us-east-1.console.aws.amazon.com/vpcconsole/home?region=us-east-1#igws:

Attach Internet Gateway

Once the internet gateway is created, select it and click the Action drop-down in the top right. Then, click Attach to VPC, and choose the VCP in the prior step.

This will give machines on that VPC access to the public internet.

#5 IAM Profile

To install GPU drivers, you need to access AWS resources like the S3 Bucket, which contains the AWS-specific Nvidia and AMD Graphics card drivers.

To do so, you need to:

  1. Create a role (if one does not exist) that contains a policy to read from S3 Buckets

  2. Assign that to any ec2 instances that need to download graphics card drivers (Note: the public drivers on vendor site do not work for AWS instances).

IAM Profile Create Role

Select AWS service and use case EC2 and hit the next button in the bottom right.

Select AmazonS3FullReadAccess and hit the Next button.

This example left the default options on the following page. Finally, hit the Create role button in the bottom right.

Once you create an instance, you need to attach that role to the instance (shown in next steps)

#6 Create Instances

Now that your VPC is configured, you are ready to start creating EC2 instances.

Access your EC2 console and select Launch Instances in the top right.

(If you already have instances created, skip this step and select Instance State -> Start if they are not running).

Row64 Server, Row64 Studio, and Desktop Dashboard Clients have different machine configuration requirements, so instances will need to be configured accordingly.

AWS has pre-configured instances, but we will go through the manual setup for creating custom instance, so you can skip the AMI section, which is for pre-configured instances.


Row64 Server - supports Ubuntu Linux 22.04 (and higher) SSD volume Type or RHEL 9.0 (and higher) SSD SSD Volume Type

Row64 Studio or dashboard clients - require access to an accelerated graphics device - Windows Server 2019 base SSD volume type (Windows Server 2022 currently does not support GPU drivers; the latest compatibility is 2019).

Architecture is x86-64


Next, select machine type.

This example chooses t3.xlarge for Row64Server (4 CPUs and 16 GB RAM) and g4ad.xlarge (AMD GPU) or g4dn.xlarge (Nvidia GPU) for Studio.

Note: Machine sizes can scale up or down depending on workload, but this is a good baseline.

Server performance scales with more cores and memory.

Next, create a key pair (or select one if it exists). Key pair is a security key file you download. It’s required to log in to your instance via SSH or RDP.

Network Settings

Next, choose network settings. It will default to the VPN and subnet you created in the previous steps.

Select the security group created in the previous steps.

Configure you storage setting. 30 GB is recommended (again, it is workload-dependent)

Do not make any changes to Advanced details.

Now, you're ready to launch you instance.

Click the Launch Instance button in the lower right corner.

#7 Remote Desktop

Now, let's connect to a running instance to Install Row64.

Hit Connect from the Instance page.


Linux Remote Desktop

Next, launch a Linux instance for Row64 server. Copy the SSH command and SSH into your Row64 server from the command line.

The key pair file should be downloaded and accessible on the machine you are running the SSH command from. Now, you're ready to start downloading and installing Row64Server.

Windows Remote Desktop

For Studio Windows Instances, once the instance is up and running, hit Connect.

Next, go to the RDP Client button.

Access the instance via Microsoft Remote Desktop client on your machine with the IP address of your instance. Then, click get password to generate a password from the key pair file from previous steps.

Keep the username: Administrator

Remote Desktop - Other Options

If Remote Desktop conflicts with your graphics, you can try the NICE remote desktop.

Open a browser and download the DCV server from their website:
https://download.nice-dcv.com/


Run the installation and accept all the defaults. The server starts automatically.

Then, return to the client machine and install the DCV client for whatever platform your running on from the same location:
https://download.nice-dcv.com/

Run the installation accept the defaults.

Use the same IP, username, and password you used for the Microsoft Remote desktop to log into the DCV Viewer on your client machine.

Graphics Driver Install

With a remote connection established to the Windows instance, we can install the GPU drivers (your Windows remote desktop will automatically drop).

Open Microsoft Powershell and follow all the instructions to download and install the AWS GPU drivers based on the instance type (AMD or Nvidia)

AMD:
https://docs.aws.amazon.com/AWSEC2/latest/WindowsGuide/install-amd-driver.html

Nvidia:
https://docs.aws.amazon.com/AWSEC2/latest/WindowsGuide/install-nvidia-driver.html#nvidia-GRID-driver

Once the installation completes, open Device Manager, select Display Adapters, right-click on Microsoft basic display adapter, and disable it.

(Otherwise, it interferes with the GPU drivers, and you would no longer be able to use Microsoft Remote desktop. This is why we installed DCV server on the machine).

Reboot the instance and reconnect via the the remote desktop. If the driver installation was successful, you should see AMD Radeon Pro (AMD) or Tesla T4 (Nvidia) in the Device Manager display adapters. In Task Manager, you should now see the GPU.

#8 Install Row64 Server

Connect to the Linux instance (Ubuntu or RHEL) through SSH, using the previous instructions.

From here, you can follow the instructions in the v3.5 release notes:
Version 3.5 Release Notes

#9 Test Server

You won’t have access to a browser on the headless server.

To test that the server is functioning properly, you can go to a desktop instance on the same private network and connect to a sample dashboard on the server provided with the setup at the following URL:

host address or hostname/dash/examples/bubble-widget

#10 Install Studio

Once your remote desktop connection to your Windows Instance is up and running, open a browser and download Row64 Studio from here:
https://app.row64.com/Downloads/

Follow the installation instructions accept the default options.

Once the installation is complete, Row64Studio will start.

Row64 Studio V3.5 should open with a pop-up to Connect to Server.

Next, enter your server's IP address or hostname. You should see the red “Not Connected” turn green “Connected”

From there, you can click browse in the top menu bar to open dashboard examples.