When HBX was launched, there was a need to be fast and agile, which was a perfect match for the cloud-based technology infrastructure services provided by Amazon Web Services (AWS).
When HBX went live, it was one of the largest production implementations of AWS in Higher Education. At that time, I was working in central IT at Harvard, where we developed a number of innovative programs to make it easier for groups across the University—such as HBX—to use services from AWS. Little did I know that I’d eventually work for HBX in supporting their Amazon implementation! All of this work has given me a great opportunity to learn and work closely with AWS as it has rapidly grown.
1. AWS adds as much computing capacity DAILY as Amazon the company had to run its entire operations as an $8.5 billion dollar enterprise in 2005.
This capacity is distributed across a huge global portfolio of datacenters. Each geographical cluster of datacenters is considered a region, and each region is comprised of multiple availability zones, which are themselves at least one—and sometimes more than one—distinct datacenters. In total there are 16 regions consisting of 42 availability zones, with another 2 regions announced and under construction.
2. AWS engineers much of their own equipment.
Given their huge scale of operations, Amazon has taken to engineering and even building many of their own components in order to optimize them for their specific needs. This has included things like hiring power engineers to rewrite how utility-sized power transformers operate, building out their own high-dentisty disk storage design, and now even manufacturing chips. For a fascinating view, watch some of the entertaining presentations from AWS Vice President James Hamilton:
3. AWS has a strong focus on rapid innovation and is frequently releasing new services and upgrades to existing services.
In order to keep this high level of development velocity, AWS services are typically owned by many small teams who are largely independent of each other. And while Amazon teams may get access to features earlier, they make use of internal services in the same way as their customers when the services are public. This ensures a high level of stability and consistency, and has been a key factor in their ability to both innovate and scale.
4. AWS runs a version of their cloud for the CIA.
While one of the early concerns about the cloud was how secure it is, this is rarely a concern anymore—even for the CIA. Running services in the cloud is often more secure than running them in your own datacenters. In fact, many AWS services are certified to comply with some of the federal government’s most stringent security requirements, and have achieved a broad range of other security certifications for health care, criminal justice, financial services data.
5. AWS is working to source 100% of its power from renewable sources.
As of the end of 2016, 45% of their power comes from renewable energy, including from four wind farms averaging over 160 Megawatts of generating capacity each. Including projects currently underway, they soon expect to have over 900 Megawatts of renewably sourced generating. One of their regions, located in Oregon, is supplied entirely by renewable sources.
6. It can still break (but it's usually less painful when it does).
As many people who were using the internet last week know—including everyone in HBX courses—even AWS can have outages that take down your services. But from having spent many years working with on-premise data centers, the frequency and severity of outages tends to be much lower.
While it is still stressful to anxiously await updates from a vendor on a cause of an outage, it is far less stressful than scrambling your own team to identify and fix a complex problem. It is also comparatively easier to build services which can operate out of multiple regions that further insulate you from down time—though it is still a large effort (and something we have yet to tackle for HBX).