Re: [Architecture] CEP - Largest Connected Component and Maximum Clique implementation

Ok. We will look into this option as well. On Thu, Aug 11, 2016 at 6:57 PM, Sajith Ravindra wrote: > Actually what I suggested was to use a graph library to keep the graph > data structure and implement these algorithms(i.e. maximum clique, largest > connected component) on

Re: [Architecture] [Dev] [IS] [Analytics] Improvement to use Siddhi streams to send notifications

On Thu, Aug 11, 2016 at 11:45 PM, Sriskandarajah Suhothayan wrote: > I think getting the claims will improve the message formatting when > sending the message, and based on the discussion with Johann they cannot > determine what claims the message formatting will need. if IS need

Re: [Architecture] [Dev] [IS] [Analytics] Improvement to use Siddhi streams to send notifications

I think getting the claims will improve the message formatting when sending the message, and based on the discussion with Johann they cannot determine what claims the message formatting will need. if IS need to send the claims it has to also read the message to understand the necessary claims.

Re: [Architecture] [Dev] [IS] [Analytics] Improvement to use Siddhi streams to send notifications

On Thu, Aug 11, 2016 at 11:32 AM, Indunil Upeksha Rathnayake < indu...@wso2.com> wrote: > Hi Suhothayan, > > You can refer [1] for the current approach we have taken in IS side when > improving notification sending with siddhi streams. As per the discussion > we had previously, this approach has

Re: [Architecture] CEP - Largest Connected Component and Maximum Clique implementation

Hi Bhagya, Though it's quite easy to implement basic graph structure/operations from the scratch, when it comes to advanced graph algorithms it might introduce some complexity code and performance. Besides, if we are planning on using this with SNA, going forward we might require other graph

Re: [Architecture] CEP - Largest Connected Component and Maximum Clique implementation

Hi Sajith, There is a possibility of doing this. However, the effort required to implement a graph is minimal. For example, a graph can be implemented using adjacency list and basic operations on a graph are very simple to implement. Regards, Bhagya Rupasinghe Software Engineer Intern Mobile :

Re: [Architecture] CEP - Largest Connected Component and Maximum Clique implementation

Hi Bhagya, Wouldn't it be easy if we use a library for the graph implementation instead of implementing your won graph implementation? IMO, if you try to implement you own graph you will have put more effort into that than the core functionality itself. Also, as we go on our requirements might

Re: [Architecture] CEP - Largest Connected Component and Maximum Clique implementation

Hi Sasikala, Input stream is not a set of vertices.The input stream contains a pair of vertices adjacent to each other. The siddhi extension accepts two adjacent vertices as the input. I have used a Hashmap to create an adjacency list to generate the graph. Regards, Bhagya Rupasinghe Software

Re: [Architecture] [MB][C5] Permission Model Revamping

Hi Manuranga, 1. I think you have misunderstood my previous mail. In our security implementation with C4 we are creating an internal role per queue. But in our C5 implementation we are hoping to avoid this and assume that the roles are created in the user store before creating the queue/topic.

Re: [Architecture] [Dev] [IS] [Analytics] Improvement to use Siddhi streams to send notifications

Hi Suhothayan, You can refer [1] for the current approach we have taken in IS side when improving notification sending with siddhi streams. As per the discussion we had previously, this approach has been taken in order to avoid the performance degradation due to the redundant loading of email