Ever wished you could upgrade your laptop by just renting a faster one in the cloud? With AWS, you can. In this guide we’ll walk through setting up a secure, budget-friendly Ubuntu development machine on AWS, complete with NICE DCV for a graphical desktop and VS Code for your day-to-day work.
Why build a cloud dev box?
- No hardware upgrades: Your old laptop becomes a thin client. The heavy lifting happens in AWS on fast CPUs and SSDs.
- Consistent environment: Your dev stack lives in the cloud—no “works on my laptop” problems.
- Always-on backups: Snapshots and EBS volumes mean you can roll back or clone your setup instantly.
- Portable & secure: If your laptop dies or gets stolen, your environment is intact in AWS.
- No local power or heat worries: The box keeps running even during power outages.
The setup we’ll build
We’ll use:
- EC2: Ubuntu 24.04 LTS instance (
t3.large, 2 vCPU, 8 GB RAM) - Private subnet (no public IP): safer and cheaper
- EC2 Instance Connect Endpoint: for private SSH access
- NICE DCV: to access a full graphical desktop remotely (free on EC2)
- VS Code: as your IDE (remote or GUI mode)
- NAT instance: tiny, used only when the dev box is running for apt/git traffic
Average monthly cost: ≈ $25–30 for 40 hours/week usage.
Step 1. VPC and networking
- Create a VPC (or reuse the default) with at least one private subnet.
- In that subnet, launch a t3.large Ubuntu 24.04 instance.
- In Advanced network settings, disable Auto-assign public IP.
- Attach a Security Group that allows:
- TCP 22 (SSH) from the VPC only
- TCP 8443 (DCV default) from the VPC only
- Create an EC2 Instance Connect Endpoint in the same subnet. This gives you browser and CLI access without any public IP.
Step 2. Outbound Internet (NAT instance)
To pull packages and git dependencies, you’ll need outbound Internet:
- Launch a t4g.nano in a public subnet and give it an Elastic IP.
- In its Security Group, allow all outbound and SSH from your home IP.
- Configure it as a NAT instance (AWS has a short guide).
- Route your private subnet’s default route (0.0.0.0/0) through this NAT instance.
- Shut it down whenever you stop your dev box—saves a few dollars a month.
Step 3. Configure the Ubuntu dev box
Connect using the AWS CLI:
aws ec2-instance-connect ssh --instance-id i-xxxxxxxxxxxx
Then install your essentials:
sudo apt update && sudo apt install -y build-essential git tmux unzip \
xauth x11-apps xfce4 dcv-server code
Enable X11 forwarding if you ever SSH directly:
sudo sed -i 's/^#X11Forwarding no/X11Forwarding yes/' /etc/ssh/sshd_config
sudo systemctl restart ssh
Set up NICE DCV:
sudo systemctl enable dcvserver --now
sudo dcv create-session dev
Now you can connect from the DCV client on your laptop to the instance’s private DNS via the EC2 Instance Connect Endpoint tunnel.
Step 4. VS Code workflow
On your laptop, install VS Code with the Remote – SSH extension.
Add to ~/.ssh/config:
Host aws-dev
HostName <private-ip-of-ec2>
ProxyCommand aws ec2-instance-connect open-tunnel --instance-id i-xxxxxxxxx --port 22
User ubuntu
Now you can open VS Code, choose “Remote SSH → aws-dev”, and work as if it’s local. The compute, builds, and tests all happen in AWS—your laptop just displays the results.
Step 5. Backups and security
- Snapshots: In the EC2 console → Volumes → Create Snapshot. Incremental, cheap, and quick to restore.
- No open ports: The box lives entirely inside the VPC. No public IPv4 = no attack surface and no $3.65/mo IP fee.
- Encryption: Use an encrypted EBS volume for peace of mind.
- Keys: Store SSH keys in AWS Secrets Manager or a password manager, not on the instance.
Step 6. Costs at a glance (40 hrs/week)
| Component | Rate | Hours | Monthly Cost (approx.) |
|---|---|---|---|
| t3.large compute | $0.0832/hr | 173 | $14.41 |
| 100 GB gp3 storage | $0.08/GB-mo | – | $8.00 |
| t4g.nano NAT instance | $0.0042/hr | 173 | $0.73 |
| Data egress (DCV ≈ 1 Mbps) | under 100 GB | – | $0 |
| Total | ≈ $23–25 / month |
That’s cheaper than a streaming-service subscription, for what’s effectively a modern 8 GB Linux workstation.
Step 7. Power off when done
When you finish a session:
aws ec2 stop-instances --instance-ids i-xxxxxxxxx i-yyyyyyyyyy # dev + NAT
Billing stops for compute instantly; storage continues (~$8/mo).
Next time, start them up and you’re right back where you left off.
The bigger picture
This setup turns your local machine into a secure terminal for a powerful, always-backed-up environment:
- No risk of theft or loss: Your code and configs never live on your laptop.
- Power-failure proof: AWS stays online through storms and travel.
- Cloneable dev boxes: Snapshot → Launch → Instant new environment for teammates.
- Predictable cost: Pay only for hours you’re actually online.
It’s a sweet spot between comfort, cost, and capability—perfect for solo devs, consultants, and anyone tired of lugging a gaming laptop just to compile code.
Would you like me to format this for publishing (e.g., Markdown with code fencing, tags, and SEO title/description), or keep it as an internal how-to guide?
Leave a Reply