Serverless Functions on Turbo Vectors
Turbo Vectors makes it possible for anyone to deploy functions to the Turbo 360 cloud and offer them for general or commercial use. This tremendously reduces the time and expense involved in building SaaS products, web apps, IoT projects, and many more situations that traditionally use client-server architecture. By removing the server piece of the architecture, Vectors enables developers to focus on custom application logic without the cumbersome overhead of server configuration (scaling, security, load balance, etc), DBA, sys admin, and other associated demands.
What Is a Vector?
Vectors are units of code that (functions) perform one task very well. For example, you may deploy a Vector that scrapes webpages for metadata found in the <head> tag such as 'title', 'description', 'image' and return the results as a JSON object. This Vector may be useful to other developers and on Turbo 360, they would be able to access it while you maintain it - imagine Github repos that can be executed from a Github interface. The Vector interface is a standard API (SDK wrappers available in the usual suspects: NPM, Javascript, Ruby Gem, etc). Vectors can be free for public use or restricted behind a paywall.
Speed, Cost, Reliability
Turbo Vectors are deployed without any server or database configuration. It offers a fully abstracted Functions-as-a-Service layer to developers for deploying as well as consuming functions. By removing all dev-ops consideration, Turbo Vectors reduces costs by billing for purely discrete usage rather than time running servers, databases, etc. In addition, the generous free tier (1,000 daily Vector calls per app) allows ample room for testing and development.
Vectors are deployed on a highly-distributed architecture with practically zero downtime. The first call of a Vector may need a short "warm up" period (1-2 seconds) but after that, each call is immediate. Contrast that with PaaS providers such as Heroku and Google App Engine which require upwards of 30 seconds on the first call to "sleeping" servers and you can immediately see the benefit of Vector deployments. After the payload is returned and the Vector completes, there are no additional server expenses incurred while the server waits for more requests - the billing unit is extremely granular.
Why Now?
As technology grows beyond the confined box of the desktop/laptop and even mobile phone, conventional client-server architecture shows cracks in its ability to handle the shifting ecosystem. Connected hardware such as smart TVs, IoT devices, drones, as well as increasing demands from machine learning technologies have made it impractical to deploy full stack web environments for every service. High levels of redundancy (millions of servers running the same backend framework with database configuration) amounts to wasteful development cycles that can be leveraged better in product-specific areas. As such, a consolidation of infrastructure is a logical progression much the way the web itself has basically coalesced around a few dominant platforms: Facebook, Google, Amazon, etc. Many businesses no longer have a dedicated website but rather, a Facebook page to promote.
What If I Want to Run My Own Architecture?
Go for it! There will always be a place for the dominant frameworks and databases and the Turbo Vector paradigm is simply another option for managing your system. It is also not mutually exclusive - Turbo 360 itself runs on a conventional AWS stack with Node servers but we also leverage Vectors for many discrete units of functionality that makes life much easier. For example, when you deploy a project on Turbo 360, the Recent Activity page is updated with your activity - that update happens through a Turbo Vector in conjunction with our REST API. As this example shows, you can use Turbo Vectors on an a-la-carte basis because the interface happens via API endpoints, similar to 3rd party SDKs and web services.
Looking Ahead
In the long run, we feel the new Vector paradigm will gain major adoption for old fashioned reasons: time and money. Vector deployments compared to conventional client-server architecture is at least an order of magnitude improvement in both areas. For now, we welcome you to test the platform and publish your own Vectors for personal and general use. We started with a few of our own basic examples HERE.
