in Book

Review: Modular Cloud Apps with OSGi

Currently, I am doing my literature studies for my master thesis about OSGi and PaaS. One of the first books I read was Modular Cloud Apps with OSGi from Paul Baker and Bert Ertman published by O’Reilly. Here is what I think about it. Maybe it is a little bit ironic sometimes, but I hope you get the point.

IMG_20140204_112527

Far more OSGi than Cloud

Modular Cloud Apps with OSGi is much more over OSGi than cloud. You will read about bundles, Bndtools, plugins and all the other (standard) OSGi stuff all over the book. The terms IaaS, PaaS and SaaS are explained on page 152 of 190 (including references and index). All stuff before is just OSGi. The only real cloud technology in the book is Amazon’s AutoScaling

Deployment

As I said, the only “cloud” chapter in the whole book is the one about deployment (~8 pages). The author explains how you can deploy nodes with Amazon’s AutoScaling and the Apache ACE provisioning server. An examples as well as some scripts are provided. And by the way, Paul Bakker, the author of the book is a contributer of Apache ACE, which brings us to another topic called Amdatu.

Amdatu

You will find one term over and over in the book: Amdatu. Amdatu is a Java OSGi framework to build (modular) cloud applications. And guess what? It is build by one of the authors of the book, Paul Bakker! This could be the reason why Amdatu seams to be the answer for everything. How should we implement authentication? With the Amdatu token provider! And what about a REST API? Yeah, Amdatu has all the libraries packed in some OSGi bundles! And ever thought about some documentation? Amdatu has already repacked the Swagger framework! And don’t forget Amdatu Social, Amdatu Mongo and all the other stuff reinventing the wheel for you while talking about modularity to help you to don’t reinvent the wheel. Also consider to not use standard frameworks like Spring, Amdatu has a more modular solution! But try to search Amdatu on Google Trends

The world is not ready for OSGi

The world is still not ready for OSGi! After reading the book you get a notion that every existing Java framework is not modular enough and the only solution is to repack them all. None shall pass.

Riding the cloud wave

The book is really good to learn something about OSGi and maybe to get some insights into SOA. You will learn what bundles are, how provisioning can be done and what REST is. But the cloud part is really weak. The book does not cover any cloud system, any cloud service or any cloud SDK/API. What a pity. Modular Apps with OSGi would have been a much better title, but cloud sells.


Edit

By the way, there is an excellent video on YouTube by the authors of the book. IMHO the video explains much better how to build cloud apps with Amdatu and OSGi than the book. Especially the part about Apache ACE becomes very clear. Take a look:




Best regards,
Thomas

  • sriram reddy

    Hi Thomas,

    I am a java developer looking for true a modular cloud platform. My requirement is to develop a modular cloud based platform with integration (ESB) and business processing(BMP). I evaluated Fuse, WSO2 and Camel with Activiti on Virgo and i considered Virgo container but i was missing right testing framework and other little bits and pieces.

    I read the book and I was very happy to know the best possibilities and various tools i needed. I developed most of my platform as per my target system requirements.

    I truly believe OSGI is the best modular java platform out there and it is ready. your right is developer is trying get a job and make a CV and follow the crowd, I recommend to follow trends . But if you want to build the application as you desire not for daily pay you don’t follow trends you look for what you need.

    To be honest I recommend this book to anyone who really want to get started with OSGI platform to start with.

  • Neil Franken

    Hi Thomas

    You pretty much sum up my feelings about the book as well. The information about OSGi is complete but nothing revolutionary. The promotion of certain frameworks above all other smacks a bit of marketing and makes you feel like you are reading an add.

    This is a nice complimentary book to read about OSGi but it misses a lot of the points. Not everyone wants to build just a normal application in the cloud. For example integration/SOA is a big field and as more system interconnect these SOA systems will need to be cloud based as well. However things like Apache Karaf is shot down as DO NOT use for new development. I find this very opinionated. Karaf and Camel play very nicely and if you are doing an integration/SOA type application it gives you the impression to avoid Karaf.

    So while this book certainly has its insights I feel the authors focus so much on their solution they dont give you a complete picture. Remove the opinions put in facts and let the reader decide.

    My 2c worth.