Hi All, Appreciate all your ideas & advices regarding $subject.
To give you some idea about EMM 1.0.0 ....... - EMM consist of two main component, MDM (Mobile Device Manage) and MAM (Mobile Application Manager). Both of them are Jaggery apps, which host on a Carbon server. - MDM developed based on IS, and MAM based on ES. So, MAM has sub components named, Store, Publisher , and MAM Console. - The artifacts related to mobile apps added via MAM will keep on a location inside the repository/deployment/.. which is need to share with all the nodes in the cluster. [*We have used svn based depsync for this in our current cluster*] - Also, MAM keep the track of mobile apps added, on registry. [*We have mount the nodes to a shared Registry*] - MAM & MDM use a separate DB for things specific for EMM, and that will use to keep the details related to policies. - MDM will dispatch a monitoring message, which will request for the current status of the devices, and the device will reply to the MDM. In this case, this message need to be dispatch by a single node, other wise there will be duplicates, and a higher traffic. Then the devices will contact the other nodes and start communicating with them. This monitoring message will dispatch with specific time intervals. [*We need to try and check this behavior when the cluster is ready*] - Admin need to access Carbon console for the tasks such as, adding tenants, assigning admin specific permissions for Admin users, etc. - MAM & MDM login to will handled via SSO. [*When it comes to cluster, we need the support of a person who have deployed SSO based Jaggery apps on a cluster. This is the point that we have stuck*] - Mobile client app will get install on mobile devices, and user needs to register to the system via that client. This is the point that the device connects to the system. Problems we are facing are......... - We have already setup a WM separated Cluster for EMM, fronted by ELB. In this case, we are currently facing an issue when redirecting to SSO login page. SSO login page will route to the 'HostName' which is the worker host, but it need to route to the 'MgtHostName'. For this, is there a way to pass the 'MgtHostName' to jaggery, with relevant port? Else can we go ahead with a cluster which has only managers? - If we go for a cluster which has only manager nodes, They all will commit to the svn simultaneously. So will it crash the svn, and how can we handle this? - Publisher will be the app use to add Mobile apps & their artifacts, and if we server publisher access via workers, then we need to give the svn & registry write access to the workers. To overcome this, is there a way to allow access to publisher app, only via Manager? What we have in mind is to route the publisher directly to manager host, and have to try it out when the cluster is ready. - Also, when it comes to tenancy we will face the issues discussed in mail thread "*[Dev] Issue with tenancy specific files in app layer*". What we currently planing to do is....... - Move to approach which has only manager nodes. - In this case, we need to find answers for svn conflict issues. Hi Shan & Team, Please add if i have missed or misspelled anything. Hi Amila/Nirmal, Thanks for the great support. Appreciate all your concerns. Thanks , Pandula *Pandula Kariyawasam K.B.* *Senior Software Engineer - QA* Mobile: +94772314510 <http://www.wso2.com>
_______________________________________________ Dev mailing list Dev@wso2.org http://wso2.org/cgi-bin/mailman/listinfo/dev
