Droplet Computing containers on Amazon WorkSpaces

Michelle Laverick – Chief Technologist

This week I spent time working with Amazon WorkSpaces. In case you don’t know, Amazon WorkSpaces is a managed, secure cloud desktop service. You can use Amazon WorkSpaces to provision either Windows or Linux desktops in just a few minutes and quickly scale to provide thousands of desktops to workers across the globe. You can pay either monthly or hourly, just for the WorkSpaces you launch, which helps you save money when compared to traditional desktops and on-premises VDI solutions. Amazon WorkSpaces helps you eliminate the complexity in managing hardware inventory, OS versions and patches, and Virtual Desktop Infrastructure (VDI), which helps simplify your desktop delivery strategy. With Amazon WorkSpaces, your users get a fast, responsive desktop of their choice that they can access anywhere, anytime, from any supported device.

In particular, I was experimenting with running Droplet Computing container software inside Amazon WorkSpaces. To make life easy I used an AWS S3 bucket to store the files I would need, including an installer (RPM and EXE), license file (.lic) and a couple of Droplet Computing Images (DCI) to play with. Droplet Computing has two types of DCI. The DCI-X offers excellent compatibility with legacy software and is designed to run on practically any hardware, whether physical or virtual. DCI-M is the latest generation of container, and is designed to run modern applications, and leverages processor attributes of Intel VT-x to offer peak performance. DCI-X is a good fit for public cloud environments because usually the cloud admin doesn’t have control over the underlying hardware or hypervisor to control whether those Intel VT-x instructions are surfaced up to the virtual machine, in same way you do with on-premises virtualized environments, such as VMware vSphere.

If you have not used S3 before it has a simple web-based UI where you drag and drop your files and select your classification of storage based on your needs and budget.

Amazon WorkSpaces is incredibly easy to set up, especially if you select a region that supports the “Quick Start” wizard. It does all the work of setting up the VPC (virtual private cloud), subnets and Directory Service requirements (you can choose from Simple Open-Source LDAP implementation, managed Active Directory or a connector into your on-premises AD infrastructure). You provision a “workspace” (either Linux or Windows-based desktop), and then allocate your users – users get a simple welcome email together with an activation code, and the requirement to change the password. Amazon WorkSpaces provide a full range of clients, and so I was pleased to find a native Apple Mac client that I could use.

I opted for the free-tier version of Linux and Windows with each allocated 2 vCPUs and 4GB of RAM with a view to using the default CPU/RAM allocations for the Droplet Computing DCI-X image which are 1 CPU/2GB of RAM. The memory demands of DCI-X are relatively modest, and so I assumed customers would want to use the smallest container to ensure they are being as efficient as possible with the consumption of cloud resources. I uploaded a copy of the DCI-X image that included Microsoft Office 2003. I find it easier in cloud environments to simply use the encapsulation that the .droplet file brings as an easy way to get an entire container build into the environment – rather than faffing around downloading installers and trying to get them into the workspace and then into the container itself. My lab environment is now all set up with different DCI’s which I’ve built over the last couple of weeks for customer demos.

My next task was to connect to Amazon WorkSpaces and, once connected, download from S3 the files needed for Droplet Computing. Then I could install the Droplet Container App.

In the Linux-based Amazon WorkSpace I was able to open a Terminal window. Droplet Computing’s RPM installer installs quickly on the command-line with sudo rpm -i

The install places the Droplet Computing Application (DCA) into the Accessories on the “MATE” graphical user interface. Amazon uses MATE rather than GNOME, as GNOME is less friendly with remote desktop protocols.

The Droplet install adds an icon to the Accessories menu in MATE, and a simple right-click adds that icon on the Amazon Linux desktop.

I’m pleased to report that DCI-X performed admirably in the context of the Amazon WorkSpaces image types I used – and I was able to get sub-second loads of applications like Word, Excel and PowerPoint from the old Office 2003 suite.

I’ve deliberately not maximized windows here, so you see the outer rim of Amazon WorkSpaces, with the Droplet Computing container within, with Microsoft Excel 2003 running inside the container. In the real world, you would maximize those windows and adjust the screen resolution in the container to optimize your screen real estate.

My Windows experience was very much the same. I was worried that Windows would consume more of the 2 vCPUs/4 GB allocations assigned than Amazon Linux would. But Amazon has done a very good job of optimizing the Windows build, so there was plenty of free RAM within which I could load the container without needing to create a bigger Windows instance.

Again, I’ve deliberately not maximized Windows here, so you see the outer rim of Amazon WorkSpaces, with the Droplet Computing container within, with Microsoft Word 2003 running inside the container. In the real world, you would maximize those windows and adjust the screen resolution in the container to optimize your screen real estate. I’ve also repositioned the Amazon WorkSpaces logo, so you see it – as normally it’s in the center of the desktop.

So, lots of flexibility here – an Apple Mac client to connect to a Linux Desktop hosted in the cloud, with a Droplet Computing container running a legacy Windows application natively – all very seamless to the end user. Because, as you know, I subscribe to the view that users don’t really give two hoots HOW applications are deployed, so long as they get their apps – so the less that the “infrastructure” gets in the way of that experience the better. And therein lies the real value of Droplet Computing – application delivery, redefined!