It all started with one computer. Then, at a point in time, we learned to connect it with other computers. Eventually, personalized computers emerged and individuals began to own the hardware that are helping them to do the work. Conversely, things run a little differently for software.
While most of us own devices, we are pretty much employing the same centralized access such as Gmail, Slack, or Dropbox and so on. Presently, there are only a few cloud providers: Amazon, Google, Microsoft and IBM and a large percentage of companies are relying on them to build or host artificial intelligence (AI) infrastructures.
To de-centralize the role of cloud
As advances continue, there is not much room within the cloud space for growth, and this becomes an advent for a new form of computing – Edge. Simplistically speaking, edge computing is to get a job done near or at the source of the data, rather than depending on massive data centers or the cloud space.
For example, virtual assistants typically sort out users’ voice commands within the cloud. Often, the speech is being processed and a compressed representation of it is being sent to the cloud. The cloud will then de-compress it and process it. Depending on the requests (i.e., asking for the weather or checking of flight schedules and so on), there may be a need to rely on other API (application program interface). All these take time.
The rumor of Amazon designing its own AI chips for Alexa has run more than a year. Indeed, if a virtual assistant could take up more processing responsibilities, it minimizes its reliance on the cloud. As a result, information gets processed faster and the overall cost is believed to be cheaper.
Specific edge computing use cases
In medicine and healthcare, smart monitoring devices, wearables, mobile applications and other telemedicine technologies that assist care providers to support remotely located patients, are likely to be the drivers of edge computing. More powerful devices could ultimately be deployed to increase the amount of collected information and provide real-time analyses and feedback.
On one hand, the quality of healthcare in isolated rural areas is reassured. Those living far from hospitals, will still be able to have the same, if not, similar level of access to medicine and healthcare as compared to other patient population. On the other hand, the messy, unstructured data are securely kept at their respective sources while only those that have been properly analyzed make it to the cloud. This may overcome the long-debated problem of having insufficient quality data to develop new medical solutions.
Similarly, such proactive, localized monitoring may also benefit the facilitation of healthcare logistics. The shipment and transportations of medications, testing equipment or even time and temperature-sensitive pharmaceuticals can be better traced from the time they depart from factories to the designated hospitals or pharmacies.
Edge computing is still at its dawn and certain domains of medicine tend to “buy early into these hypes” but takes a long time for real adoption. Let’s hope it will not take long.
A science writer with data background and an interest in the current affair, culture, and arts; a no-med from an (almost) all-med family. Follow on Twitter.