This repo was created as a demonstration for a talk (to be) given at MongoDB World 2017 by Michael Lynn
Let's break the term down...
We know that MongoDB is a database software platform based on NoSQL.
This phrase means that whatever words precede will be vended in a an on-demand manner. There are several characteristics (requirements) of an "as a service" offering.
On-demand
To be an as-a-service offering, the product or service should be available to customers immediately (or close) to their demand for said product or service.
This can represent a major challenge when talking about databases. In order to be able to provide on-demand service, you may have to pre-provision hardware, or develop some mechanism to enable you to deploy hardware and operating systems in an extremely rapid fashion.
Broad Access
Consumer has access across a variety of networks / devices.
Resource Pooling
Provider compute resources are shared across many requesters.
Rapid Elasticity
Compute resources can be scaled rapidly outward/inward commensurate with demand.
Measured Service
Resource utilization is monitored, controlled and reported.
There are countless ways to implement Database as a Service. What's contained in this repository represents one approach... maybe not even the most efficient, or effective method. In fact, as an employee of MongoDB, I feel compelled to mention the fact that this problem has been solved... not only solved - but solved in an extremely effective, and efficient manner. You can use this service to satisfy your own MongoDB as a Service offering. This may not give you the flexibility or control you require to vend this service to your internal users. THAT is precisely why this talk, and this repository exist.
NoSQL solutions address a very import gap that existed in the market. MongoDB was created to specifically address the gap in features that was created by the changing technology world. Prior to NoSQL and MongoDB, the data world was dominated by Relational Database Management Systems (RDBMS).
RDBMS systems were created during a time when computing resources were extremely expensive. Disk, memory and CPU were massively expensive, especially in relation to the compensation we paid to the developers and dba's responsible for developing, and managing this data.
MongoDB was created to address the gap between the world of relational database technologies and the world we live in today - where the volume, velocity and variety of data is quite different than when RDBMS was created.
In fact, you should. Stop reading this and go check out MongoDB Atlas. However, if after checking out Atlas, you find that you require greater control over the service you vend to your internal customers, perhaps this repo and the information contained herein will be interesting.
Flexibility
| RDBMS | MongoDB |
|---|---|
| RDBMS require a rigid schema defined in advance and must be maintained in lock-step with the application. | MongoDB supports a schemaless approach to application design. |
Scalability
| RDBMS | MongoDB |
|---|---|
| Most relational solutions approach scalability in terms of veritcal scale - forcing you to get larger and larger server infrastructure to support your increasing demands. | MongoDB supports horizontal scale through implementation of database partitioning - known as sharding. |
Availability
| RDBMS | MongoDB |
|---|---|
| Achieving high availability with relational technologies typically involves a number of third party packages to support clustering of resources. This complicates the operational model. | MongoDB was built from the ground up to support high availbility through implementation of database replica copies - we refer to these as replica-sets. |
If you want to create your own MongoDB as a Service, you may use this project as a starting point. You will need the following in order to proceed:
If you intend to start building your own MongoDB as a Service offering to your internal customers, it will be necessary for you to have the following components in place.
- Knowledge of the MEAN Stack
This project was written using NodeJS, ExpressJS and MongoDB.
- An AWS Account
Because I can't know what type of hardware you have available, I chose to create this demo system leveraging AWS, more specifically, EC2 instances.
- An AWS Key ID, and secret access key
You need to set up your AWS security credentials before the sample code is able
to connect to AWS. You can do this by creating a file named "credentials" at ~/.aws/
(C:\Users\USER_NAME\.aws\ for Windows users) and saving the following lines in the file:
[default]
aws_access_key_id = <your access key id>
aws_secret_access_key = <your secret key>
See the Security Credentials page
for more information on getting your keys. For more information on configuring boto3,
check out the Quickstart section in the developer guide.
A step by step series of examples that tell you have to get a development env running. Before installing MDBaaS, be aware that you must have Ansible installed.
$ git clone git://github.com/mrlynn/mdbaas.git
$ npm install
$ cp .env.example .env
$ vi .env
$ cp config/config.example.js config/config.js
$ vi config/config.js
$ mkdir ~/.aws
$ echo "[default]" > ~/.aws/credentials
$ echo "aws_access_key_id = ONOTAREALKEYART2NOTAREALKEY" >> ~/.aws/credentials
$ echo "aws_secret_access_key = +1n4Mo99t93reallyYcejkeyiix4BLika++7KjFoVZ" >> ~/.aws/credentials
$ npm startThen visit http://localhost:3000/
The system is designed to show how you may build an internal MongoDB as a service offering in your company. I've created a sample set of MongoDB services based on t-shirt sizes. In order to create these examples, you should have MongoDB running on the same system where you installed the code from this repo.
Next, change to the project directory (at the same level as app.js) and type:
$ node data/create-products.jsThe cursor should just return after a second or two. This nodeJS script will create the following products for your demo catalog:
| Product | Description |
|---|---|
| Small | Small MongoDB Instance |
| Medium | Medium MongoDB Instance |
| Large | Large MongoDB Instance |
| Extra Lg | Extra Lg MongoDB Instance |
| OpsManager | Ops Manager Instance |
- TBD
Please read CONTRIBUTING.md for details on our code of conduct, and the process for submitting pull requests to us.
This project is licensed under the MIT License - see the LICENSE.md file for details
- TBD