Training to go to space means spending a lot of time learning what to do when things don’t go the way that they should. Recently, for example, I flew a simulation in which a power failure occurred at a very inopportune moment, and I and the rest of the crew had to scramble to restore power to the vehicle’s most critical systems. The day before that, I flew on another simulation in which the capsule’s sequencer (one of the onboard computers) failed during reentry. This meant that among other things, the parachutes did not deploy as they should, and had to be manually deployed. Being that these types of failures can quickly lead to a life-threatening situation, the crew spends a lot of time learning what to do when things go wrong. As exciting as it might be to put on a spacesuit and get strapped into a capsule for a simulated mission, I first had to spend a lot of time in a classroom learning about the spacecraft’s systems, abort procedures, and how to climb into a life raft while you are wearing a spacesuit (it’s a lot harder than you might think). As strange as it sounds, one of the exercises that I worked through in such a class completely changed the way that I think about technology, especially cloud services. Let me explain.
One of the big problems with going into space is that you don’t have the luxury of bringing along frivolous items. There is a limit to the amount of weight that the vehicle can carry, plus there is a huge cost associated with everything that you bring along. By some estimates, sending items into space can cost up to $10,000 per pound. Besides that, the capsule is a cramped environment without a lot of extra room. Much of the capsule’s available space is taken up by science experiments, food, and other necessities, so you have to be extremely selective about any other items that you bring along.
Given these and other limitations, my crewmates and I were required to work through an exercise in which we had to decide which types of survival supplies to bring along on a simulated space mission. We could bring anything that we wanted, but we had to justify each item’s cost, weight, and the amount of space that it occupied inside the capsule.
A simple assignment? Not quite
On the surface, this probably seems like a fairly simple assignment. However, the mission parameters play a huge role in the item selection process. For example, the simulated mission was to launch from Florida into an Earth orbit with an orbital inclination of 51.6 degrees (which matches the orbit of the International Space Station). This means that the spacecraft’s ground track will always fall between 51.6 degrees north latitude and 51.6 degrees south latitude. As such, polar survival gear is unnecessary because the spacecraft will never pass over a polar region.
Another, perhaps even more important consideration is that launching from Florida to the International Space Station means launching prograde over water. Since critical failures are statistically most likely to occur during launch and reentry, water survival gear was an absolute must. Sure, the capsule is designed to land in water, and is therefore able to float, but there are situations that can force you to leave the capsule before the recovery crew arrives. These situations might include a fire, a nitrogen tetroxide leak, or even more benign situations such as the life support batteries going dead a couple of hours after landing.
As I and the other crew members worked through the exercise of deciding what survival supplies to bring along on the simulated space mission, we placed a high priority on sea survival gear. Of course, there was also a possibility that in an emergency, the capsule could come down on dry land. This possibility led to some lengthy debates about whether or not we had the luxury of bringing various survival items on board. Ultimately, we based our decisions around the statistical probability that an item would be used, and on the potential consequences of not having that item. This meant that items such as a first aid kit and a flashlight were included among the mission’s supplies, while items such as parkas and solar chargers didn’t make the cut (a parka is unnecessary because a spacesuit will keep you warm).
At dinner that night, the group’s conversation ended up gravitating back to the simulation that we had completed earlier in the day. Someone made the point that because our item selection had been so limited by weight and volume that it would probably be a good idea to select multipurpose items whenever possible. By that logic, a Swiss Army knife might for example be preferable to a Rambo-style jungle survival knife.
Believe it or not, that discussion got me thinking about the current state of technology, especially the smartphone. After all, the smartphones that we all carry around can replace countless devices such as a wrist watch, a camera, a video camera, a GPS device, a compass, a flashlight, and the list goes on and on.
Single-purpose vs. multipurpose
Even so, there is a huge difference between having multipurpose functionality and best-of-breed functionality. Last year, for example, I had to do a project that involved spending a considerable amount of time in the wilderness. Because there was a very real possibility of getting seriously lost, I brought along two different navigation devices — a GPS enabled smartphone, and a Garmin Montana 680T. Although either device would have gotten me back home, I noticed that the Garmin device was consistently more accurate than the GPS maps on my smartphone.
In a way this makes perfect sense. The Garmin device costs just as much as some smartphones, and yet has been engineered for a single purpose. It does one thing, and does it well.
While most of us probably aren’t going to find ourselves in the wilderness, miles from the nearest road, depending solely on an electronic device to keep us from becoming hopelessly lost, my lessons learned with regard to the smartphone are also applicable to the world of cloud services.
In a lot of ways, the big cloud providers are a lot like smartphone manufacturers. Think about it for a moment. Whenever the latest smartphone hits the market, it is a safe bet that it will be jam-packed with as much gadgetry as the manufacturer can cram into the phone’s case. Similarly, cloud providers such as Amazon and Microsoft give their subscribers access to dozens of different services.
When a cloud provider (whoever that provider might be) places their emphasis on creating a vast array of services and features, there is a good possibility that at least some of those services and features will be good enough to get by on, but will be far from being industry-leading solutions. And in some cases this might be OK. Think about smartphones, for example. A smartphone’s camera is never going to produce the same image quality as a high-end digital SLR camera, but for day-to-day use a smartphone camera is perfectly adequate. Similarly, a cloud provider’s load balancer might be good enough, even if it does not have all the features of an industry-leading solution.
In other cases however, it probably makes more sense to avoid using a cloud provider’s solution and instead take a multicloud approach (use a competing service that is hosted by a different cloud provider) or to purchase a third-party solution and run it in a cloud-based virtual machine. This can be especially true for security-related tools and services, given the critical nature of keeping your resources secure.
Bottom line, some cloud services are better than others
When it comes to cloud services, you cannot automatically assume that every service that is offered by a provider is going to be top-notch. Some cloud services will inevitably be inferior to (or far more costly than) competing services and solutions. That is the main reason why the multicloud approach has become so popular over the last few years. A multicloud environment gives you the ability to capitalize on each provider’s strengths, while avoiding services that are sub par.
It is also worth noting that you aren’t limited solely to using those services provided by the cloud providers. The larger cloud providers offer a marketplace within which third-party vendors offer solutions that have been architected to run in the cloud. Although these marketplace items tend to cost more than the basic cloud services, they have been specifically engineered for a single purpose and usually do a far better job than the basic services offered by the cloud provider.
Featured image: Shutterstock