Serverless computing or function server based computing has become popular recently for several reasons. One, function servers don’t require the developer to provision or configure a server. Another reason, for many applications the overall performance is better, faster execution time. One of the more compelling reasons is that serverless computing costs less than server based computing. There are many other reasons to go serverless like excellent scalability. Of course, there are also reasons to not go serverless. Function servers are designed to operate with applications that are small in size, often consisting of just one to 1000 simple functions. For apps that must process large workloads, specifically workloads that must process large amounts of data, its hard to make a a case for serverless. On the other hand, for Internet of Things (IoT) applications, where data from active sensors is processed by just calling one or a few math functions, function servers make the most sense.
No-Code Function Servers
From a development point of view, function servers are a light-weight code approach. That is creating an app for a function server takes much less time, work and coding skill to develop than developing one that runs on a server. A lot of the middleware, that is the server software, and numerous server innuendos no longer have to be considered. More so, for the client, the one who uses the function sever driven app, the app can be accessed without a GUI interface, that is the app can be directly queried with a query statement in the URL browser bar.
The Google Math Function Server
One of the most useful function servers can be accessed easily from your Google search box. To access the Google math function server, all one has to do is type in the math function you want calculated. Google will not only take you to its calculator but press all the calculator buttons for you. In the example below, the math expression “8log100” was typed into the search box, This resulted in the correct answer “16” appearing in the calculator’s answer box. Try it for yourself,, paste 10+1000*(0.233) in the Google search bar and you will quickly see the answer is 243, no button pushing required.
Math Function Server, Example One: Calculate the value of the mathematical expression, 8*log(100) with Google’s Math Function Server
Math Function Server Solution One: Finding what the logarithm of 100 is, multiplied by 8, requires that you enter, 8log100, into the Google search bar, as shown below. The function server assumes multiplication without a multiplication sign.
Google’s Math Function Server responds immediately. It takes you to the calculator page which has already computed the answer, 16, as shown below.
As one can see, accessing a function server makes it easier for the client to access an app, in this case Googles math function app. As a developer, the next question would be how to access Googles math function server to execute a math function list that processes a list of independent variables that is to feed into a list of functions. That is call the Google Math Function Server within the app and drive it with functions and independent variable data.
Function Processors:
One may wonder if serverless is a misnomer and that perhaps a function server is a more-straight forward name.. It would seem that all apps would require access to a server. That may be so, but in reality what an app needs access to is some sort of processor. It doesn’t necessarily need to be a 256-core, 64-bit server processor, an old fashioned single-core 8-bit microprocessor.could easily work. The ideal server-based function processor may very well be an Input/Output ( IO) intensive design that could service tens of thousands of apps that just need to do a few calculations each (that is a few functions each). Function servers are designed to service millions of asynchronous interrupts without latency. Often heat, for example, must be turned off or on immediately, that is once a sensor, through an interrupt, indicates the temperature is too hot or cold.
Graphing functions, or at least making the calculations needed to plot the graph of a function is not out of the application scope of a function processor. However in real-life it may just be as efficient, or more so, to leave that task to the client side’s microprocessor. The equations listed in the AlgebraMatrix below could be used in the apps function list and any number of client side or server side plotting modules could be used.
