Paul Fremantle: going from agile business to adaptive business
By Patrick Toenz, Managing Director Yenlo Switzerland
Paul Freemantle – WSO2 cofounder – presented a very interesting keynote. He has had a 2-year break from his work as CTO, as he was working on his PhD in the Internet of Things. This might have contributed to the top-down view he presented today. The enthusiasm in his presentation was contagious, as he made me look at WSO2 from yet another and new perspective. He translated the famous “agile business” into the more concrete “adaptive business” and described how to make this work in real life. To quickly adapt to new demand, workload changes, and therefore to new business capabilities and costs, a permanent adaptation process must be implemented. Or, as Paul put it: “You are not done with a onetime exercise”.
Fortunately, Paul had some good news for us too. Becoming an adaptive business only requires two steps:
- Simplification: put an abstraction layer onto your enterprise applications’ functionality
- Governance: implement a process to manage the complexity of these abstractions
Sounds too simple? Well, it might need some work to get there. Business capabilities and software features need to be provided as APIs, Events or Streams (or a combination of all three) – depending on the use case. So, an architecture and design process with a clear focus on integration/interfaces should drive each new business and or software venture. This way, you get a proper abstraction layer that is required for simplification and adaptability. Second, you need to implement that famous governance that covers of course planning, change and version control, but also publication of the available APIs, access control and monitoring.
Paul presented a very concise analysis with the – not too surprising – conclusion that WSO2 is the only open source company that can provide all of the required components for that journey. I agree with his presentation and analysis – with one exception: yes, the runtime can be solved with WSO2 modules – e.g. API-Management, ESB, Identity Management, Analytics etc. However for managing the complexity in planning and execution, we need another management tool: a comprehensive and systematically used Enterprise Architecture Management (EAM) repository. In theory, you could also build this with the WSO2 Governance Registry module. In the light of the many failed integration exercises by all industry “gorillas”, it might actually be a good idea to keep strategic planning separate from implementation.
Crafting an API Marketplace
By Rob Blaauboer, Integration consultant & Trainer
An API Marketplace makes APIs manageable, monitorable and monetizable. We were given the example of Dialog (Asia based Telco), that worked on a new exchange platform. As can be expected, it takes a lot of effort to create such a model, from a technical perspective as well as from a user perspective. Especially in large organizations, there might be resistance to using APIs that have been developed by other units. The success factors in the story of Dialog were therefore gamification and incentives for publishing and the use of APIs. Another thing we learnt is that, depending on the enterprise, there might be an aggregated marketplace from a number of business unit marketplaces.
Dance of the Ballerina
By Rob Blaauboer, Integration consultant & Trainer
The reason for a new language, now called Ballerina, is that WSO2 wants to make it easier to consume and produce networked services and applications. It’s basically an evolution of the best aspects of many languages. Second, Ballerina will make it easier to handle parallel processing, asynchronous programming, deal with hardware and other failures and also allow the use of microservices and serverless setups. Ballerina knows JSON, XML and data and supports transactional modes which is a departure from, for instance, the current WSO2 ESB.
There are neat new tricks that will allow XML to be defined as pure XML, not strings that need to be parsed. Second, SQL will be supported from Ballerina including streaming data tables. The query will turn into a JSON variable directly. Ballerina also knows popular APIs, like Facebook, Twitter, Gmail, LinkedIn, and so on. There will be a store-like concept most likely for connectors, like the store we now have. Of course, Ballerina supports extensible authentication since servers need security. And obviously, swagger can be imported as well since ballerina services interfaces is actually Swagger. Ballerina can do parallel processing natively which can be extremely powerful in the right setup. Ballerina is still under development (v0.95) but actually looks quite powerful but will be the biggest paradigm shift for WSO2 of the last couple of years. Nothing but good news in this field!
The WSO2 Stream Processor
By Nils Eckert, Integration Consultant
WSO2 introduced the next generation of their analytics platform, named WSO2 Stream Processor. This platform revolves around the Complex Event Processor, Incremental Time Series Aggregation and Machine Learning. In addition, WSO2 will provide out of the box solutions built on top of this WSO2 Stream Processor, for example, Financial and Banking Analytics or Retail Analytics.
The WSO2 Stream Processor is meant to be lightweight, lean and cloud native. It can be used in a minimum HA deployment with only two nodes or as a distributed deployment with Kafka. The minimum HA deployment can support about 100K events per second and is expected to be sufficient in most situations. The main artefact of the WSO2 Stream Processor is called a siddhi-app and is using the same syntax as the execution plan in DAS as of today. Batch analytics via Spark have been removed. However, most of what is done in DAS via Spark scripts can also be done within a siddhi app. In addition, receivers, publishers, streams and data sinks are no longer separate artefacts. Instead, they need to be defined as annotations in the siddhi-app.
The WSO2 Stream Processor supports all the following patterns:
– Streaming data pre-processing
– Data store integration
– Streaming data summarization
– KPI analysis and alerts
– Event correlation and trend analysis
– Real-time learning and predictions (machine learning)
The WSO2 Stream Processor also provides support for custom dashboards, like what is currently available in DAS. In addition, it provides tools for development and monitoring (apps, nodes and clusters).
The WSO2 Stream Processor is expected to be released in Q1 in 2018. So only a couple of months to go!
Can’t get enough?
Good. Neither can we. This is why we can’t wait for tomorrow, when our client FOITT will tell about Full Stack Automation, Pitfalls and Solutions. Also tomorrow, our very own colleague Rob Blaauboer will attend the panel discussion Case Studies From WSO2 Partners: Engaging for Delivery. As for Wednesday, our customer HHNK will present The Win-Win-Win of Water Authority HHNK.
We’ll tell you all about it, but not before we’ve attended tonight’s networking event and had a good night’s sleep. Cheers and talk to you soon!
Feeling bad that you’re not in London? Not to worry. Thanks to our whitepaper Go Digital, you can still work on the digital transformation of your own business. Fine, it may not be the same as attending the WSO2 Con in London, but it’s still a great start!