These are the news items I've curated in my monitoring of the API space that have some relevance to the API definition conversation and I wanted to include in my research. I'm using all of these links to better understand how the space is testing their APIs, going beyond just monitoring and understand the details of each request and response.10 Apr 2015
I've been monitoring an emerging slice of the API, that has been dubbed "microservices" for some time now, and you've even heard me explore its use, when describing my architectural approach to redesigning more core API stack for my internal systems. I’m slowly redesigning my internal API stack, keeping them as small as possible, throughout all aspects of operations, deploying each endpoint in a single dockerized container, that contains the OS, database, web server and all server side API code. I'm also using Github, in conjunction with APIs.json to assist me in my orchestration, throughout every aspect of these APIs lifecycle from design to testing—all a new approach for me when managing my APIs.
Is defining my APIs in this way, so that the definition is much more than just the actual API endpoints, but also include the entire backend, and lifecycle workflow Microservices? Fuck if I know. One thing I do know, is I've had enough conversations with folks who are doing microservices, resulting in me saying several times that the micro services definition is something very personal. Something which the first 10 times I said , sounded very positive, but after experiencing multiple people telling me what I'm doing isn't microservices, I can tell the term definitely will continue to be very personal—very much in the same way conversations around REST or Hypermedia has gotten personal on forums, and HN threads.
You know what this tells me about microservices? Is it is more about power, control and influence, everything from enterprise architects stating their endorsement of the concept, down to the individual architects, who love to make sure you know that the way you are doing it, is WRONG! I don't think there is a single definition of "microservice", and much like REST, I think we'll hear plenty of “right” definitions, by the enterprise justifying what they are doing, and those who are selling something to the enterprise. Is there anything wrong with this? Nope. Just not my style, not my game, and has shown me over the couple months how the concept is not a fit for me.
I’m going to stick in the realm of API. It is a completely bullshit term, that means everything, but you know what? Somehow it escaped the IT, architect, and vendor ownership that SOA and now microservices possesses. For me APIs have long been more than just the tech, it is also about the business and politics of any platform implementation, so why can’t it contain my architecture styles as well? So from here forward I will be closely paying attention to microservices conversations, but you won’t hear me use the word, as I’m more comfortable in API land, and what it has come to mean to the wider public—I just don’t give a shit about what the enterprise adopts, or doesn't.
Microservices will reflect the power and control of the entities behind them, unlike APIs, which also possesses that characteristic, but also quickly shifts the conversation to be more about the innovation and opportunity that developers bring to the table, and the end-users who put that to use. Let’s make sure as we do not get caught up in the architectural discussions behind the technical curtain, and that we remember why all of this API stuff is working. Let's not get caught up in all the value being just about our architectural styles, and re-live a classic IT tale, ignoring that the real value is what people actually do with our services.
The Enterprise Will Make The Same Mistakes With API And Microservices That They Did With SOA, Because Essential API Concepts Go Right Over Their Head13 Mar 2015
I would say the enterprise space fleet has successfully shifted their course, heading in the general direction of everything API. SAP, Oracle, IBM, Microsoft, and the rest of the usual suspects have pledged their faith to APIs, mobile, and IoT, all in a very public way. For those of us who jumped off the SOA bandwagon a while back (getting on the API bandwagon, yeehaw), this can be amusing to watch, cause the API bandwagon is better y'know?. ;-)
I sincerely hope that these large companies can make a shift towards an API way of life, or a microservice devops way of operating, or whatever the cool enterprise kids are calling it these days. Some companies will find success in these areas, while many, many others will just talk about it, and buy a wide range of vendor services that will help them talk about it—there will be a lot of money to be made (and spent).
The thing I fear, is that many of the core principles of why this whole thing is working will go right over the heads of many technologies and business leaders in the enterprise. Elements like simplicity, transparency, openness, and functioning in a decoupled, decentralized way. You are the enterprise, many of the essential ingredients of API, are just fundamentally at odds with what you are. You are a big, complex, powerful, and political corporate presence—a beast that does not like being taken apart, reinvented over and over, and being forced to work in concert with other internal, partner, and public resources, in an open way.
This doesn't mean that some enterprise entities won’t be successful—some will. I will be carefully watching the waves of enterprise organizations, for these tech savvy few, who internalize the core philosophy around APIs, successfully uploading their corporate soul to the API singularity. However, I predict I will have to work very hard to find these companies amidst the wreckage of the enterprise organizations that make the same mistakes they did with SOA, because APIs are just not compatible with their DNA, and ultimately will be rejected for reasons that are more enterprise than they are API.
I'm just getting started exploring the ways to use APIs.json when it comes organize my new Docker fueled, micro services stack. I’ve been using APIs.son to describe each micro service, as well as define the overall collection of almost 20 micro services. I’m using the include collection, as a navigation for the loosely coupled stack of micro-services, and my friend Chris Spiliotopoulos (@chefarchitect), has done some more work with his Swagger.ed, delivering some APIs.json goodness that is in alignment with where I want to take all of this.
Adding to the discovery and visualization work he did for Swagger, Chris enabled Swagger.ed to look for valid APIs.son files as well, so when you browse to any APIs.son, like the one I have for my micro services stack, you see the little APIs.son search icon in the address bar.
When you click on the APIs.son search icon in the address bar, you get a very cool visualization. Its pretty basic at the moment, just a visual catalog of the APIs available in the include collection of my stack, but when you connect up with the Swagger visualization work he's already done, we could have a pretty cool API catalog for managing and exploring microservices.
I have almost 20 micro-services listed, and Swagger.ed gives me the ability to navigate it in a very interactive way. Whats next? We don’t know…it is about exploration, and finding out the most meaningful way of exploring the APIs I deploy and aggregate into APIs.son collections.
I can see having a visual catalog of all of my API design that I collect, then deploy, and evolve them as needed for various parts of my infrastructure or for my clients, into other sub-collections? Disconnected collections? Loosely coupled collections? Not sure how tight I want things, part of the micro service definition for me is the size of the network connecting the services, as well as the services theselves.
API Management Infrastructure And Service Composition Is Key To Orchestration With Microservices In A Containerized World02 Feb 2015
As I work to redefine my world using microservices, I have this sudden realization how important my API management infrastructure is to all of this. Each one of my microservices are little APIs that do one thing, and do it well, relying on my API management infrastructure to know who should be accessing, and exactly how much of the resource they should have access to.
My note API shouldn’t have to know anything about my users, it is just trained to ask my API management infrastructure, if each user has proper credentials to accessing the resource, and what the service composition will allow them to do with it (aka read, write, how much, etc.) My note API does what it does best, store notes, and relies on my API management layer to do what it does best--manage access to the microservice.
This approach to API management has llowed me to deploy any number of microservices, using my API management infrastructure to compose my various service packages—this is called service composition. I employ 3Scale infrastructure for all my API / microservice management, which I use to define different service tiers like retail, wholesale, internal, and other service specific groupings. When users sign up for API access, I add them to one of the service tiers, and my API service composition layer handles the rest.
Modern API management service composition is the magic hand-waiving in my microservice orchestration, and without it, it would be much more work for me to compose using microservices in this containerized API world that is unfolding.
Disclosure: 3Scale is an API Evangelist partner.
I'm using APIs.json to organize my swagger defined microservices running in docker containers, and using the machine readable API index to drive navigation between microservices organized in a single collection. APIs.json provides a simple, machine readable way to index the technology, business, and political elements of each microservice I deploy.
As I was auditing the 18+ microservices I’ve setup for my core operations, I wanted to audit each one, and make sure I had included various elements in each APIs definition, like where developers can onboard, terms of service, and where to find related code samples. To accomplish this, I was able to just compare each microservice APIs.json, with a master APIs.json template I had established while planning my original microservice stack.
Sometimes it is just as important to know what is missing, right alongside what is available for all of my microservices, and APIs.json is proving to be a simple way of understanding this layer of my API operation at scale.
If you think there is a link I should have listed here feel free to tweet it at me, or submit as a Github issue. Even though I do this full time, I'm still a one person show, and I miss quite a bit, and depend on my network to help me know what is going on.