Hey Tony , thanks for the explanation. It did clear out many questions ! As far as i can understand we are calling performers as microservice here some of which are implemented already like stream, virtual_sensor and zmq. What other mircroservice needs to be implemented ? And if you can attach the link to your presentation that would be great ! Thanks !
On Fri, Oct 21, 2016 at 12:45 AM, Tony Faustini <[email protected]> wrote: > Hi Shivansh, good questions. What is the road map and how does iota (and > Akka clustering fit in). IoT is not simply the collection of large numbers > of devices on the internet - we have had that for decades in one form or > another. IoT is about how all these devices can interact with each other > and with the people they work with. So there is a human side to IoT which > we can talk about later for now let’s focus on what needs to happen in the > “backend” of an IoT system. the iota Fey engine focuses of the micro > services that are what people and things needs to perform useful activity. > A simple example is “if a temperature sensor is greater than 30 C. then > send me an Email”. The actions that are the result of the state of a device > or set of devices has one or more actions associated with it. These actions > are the use case for the Fey engine and it’s performers (Akka actors). Now > imagine you are dealing with very large numbers of users, an even larger > numbers of IoT devices and with many actions continuously running. You need > a scalable fault tolerant architecture to deal with such a volume of > actions (micros-services). This is the role we envision for the Fey engine. > This together with an ever growing number of performers (Akka actors) > creates a rich eco-system of micro-services. Having Fey run on a single > machine work for a certain volume of micro-services but once the volume > gets large the resources of a single machine is inadequate. You now need a > scalable micro-services architecture. This is why we need an Akka cluster. > Having that Akka cluster run on a SMACK (Spark, Mesos, Akka, Cassandra, and > Kafka) stack is a good way to scale such a solution. So building Fey on > Akka is important and extending this to Akka clustering is key. Having all > this run on a Mesos platform has other benefits that we can also discuss. > Another question we might ask ourselves is does Fey need a Mesos Framework > like say Spark has? I think Barbara has developed such a framework if it is > needed and she is able to contribute it. > > Shivansh, this only describes the micro-services part of the IoT system. > Perhaps it’s best to focus on this for the moment so we can get an initial > version of this done in the next few months. We can then extend this to > include a broader vision of the IoT platform (like the one I presented at > Apache Con earlier this year in Vancouver) > > Thanks > -Tony > > > On Oct 20, 2016, at 10:45 AM, Shivansh Srivastava <[email protected]> > wrote: > > Thanks for the reply Tony ! > And its Shivansh and not Srinivas ;) > > We have some experience in Akka Cluster ! We can make that work ! But can > you please let me know a whole overview , what is the use case that we are > trying to solve ! What is the exact idea behind this project ! I know the > part of making a open source platform for IoT and fey-core. I am just not > able to map these things properly. If you can help me out that would be > great ! > > Thanks > > On Thu, Oct 20, 2016 at 10:27 PM, Tony Faustini <[email protected]> wrote: > >> Srinivas and Hadrian, yes we need to accelerate the work on iota. What do >> we have today? >> >> i) The Fey core component (the engine for iota) >> ii) A number of components that when wired together (using the iota >> engine) are used to build micro services >> iii) documentation included inside the components - typically in the >> README.md files This is true for the core engine and for a number of >> components. >> >> Suggestions for those that want to contribute to the project >> >> 1) Write more documented components so that more micro-services can be >> build. I have a number of components that I will add to this. Could this >> start with components that connect to other Apache based components? We can >> include other technologies outside of Apache if we can include those >> technologies within an Apache project. >> 2) Making iota able to run in Akka cluster mode - today’s engine doesn’t >> do that today it runs on a single machine. We need Akka clustering to >> enable additional scalability/redundancy beyond a single machine. >> 3) What would be the best way of running iota on Apache Mesos and how >> does it relate to 2) ? >> 4) Write better documentation - it’s one thing to have the documentation >> for components and the engine. What other documents are needed and who will >> setup to write them? >> >> These are a few thoughts that I have >> >> Thanks for moving this project forward I look forward to hearing what is >> said on these points and ideas that others have. >> >> -Tony >> >> >> >> >> >> On Oct 20, 2016, at 4:43 AM, Shivansh Srivastava <[email protected]> >> wrote: >> >> I know there are quite a difference between Beam and IotA , and I am not >> comparing the implementation and the use case that they are trying to >> solve. All i am saying is that it started just a month after IoTa , and is >> constantly being getting better day by day. But the development in IoTa >> seems slow. How can we make that fast ? if we can get a better insight in >> what are the stories that we have to implement. ? When are we planning the >> first release? >> >> On Thu, Oct 20, 2016 at 4:45 PM, Hadrian Zbarcea <[email protected]> >> wrote: >> >>> The first step is exactly what you're doing. Propose improvements and >>> keep contributing. >>> >>> There are many differences between Beam and Iota, we can discuss them if >>> you want, but this is probably not the right forum. >>> >>> Cheers, >>> Hadrian >>> >>> >>> On 10/20/2016 02:15 AM, Shivansh Srivastava wrote: >>> >>>> Hey All, >>>> >>>> i really want to contribute to this project , but the development seems >>>> to be very slow. >>>> The project started on*2016-01-20 *at the same time as the project >>>> *Apache-beam* has started dated: >>>> *2016-02-01. >>>> >>>> * >>>> And Apache beam is continously getting better and better each day, but i >>>> do not see that much activity on IoTA. You can see the Commit Graph over >>>> the period and compare : >>>> * >>>> * >>>> *Apache Beam: >>>> >>>> >>>> * >>>> *Apache IoTA: >>>> >>>> >>>> >>>> >>>> * >>>> I am just concerned where are we headed? >>>> >>>> We have two to three members who are really interested in contributing >>>> to this project. Hence if someone can show us the right path in which we >>>> can contribute that would be great! >>>> Also if someone can*let us know the procedure of becoming the Apache >>>> IotA Commitor* that would be perfect ! >>>> >>>> Thanks >>>> -- >>>> /Best Regards | Shivansh/ >>>> /Software Consultant// / >>>> */Knoldus Software LLP/* >>>> >>>> /India - US - Canada// >>>> Twitter <http://www.twitter.com/shiv4nsh> | FB >>>> <http://www.facebook.com/xeruo> | LinkedIn >>>> <https://in.linkedin.com/pub/shivansh-srivastava/3a/204/251>/ >>>> >>> >> >> >> -- >> *Best Regards | Shivansh* >> *Software Consultant* >> *Knoldus Software LLP* >> >> *India - US - Canada* >> * Twitter <http://www.twitter.com/shiv4nsh> | FB >> <http://www.facebook.com/xeruo> | LinkedIn >> <https://in.linkedin.com/pub/shivansh-srivastava/3a/204/251>* >> >> >> > > > -- > *Best Regards | Shivansh* > *Software Consultant* > *Knoldus Software LLP* > > *India - US - Canada* > * Twitter <http://www.twitter.com/shiv4nsh> | FB > <http://www.facebook.com/xeruo> | LinkedIn > <https://in.linkedin.com/pub/shivansh-srivastava/3a/204/251>* > > > -- *Best Regards | Shivansh* *Software Consultant* *Knoldus Software LLP* *India - US - Canada* * Twitter <http://www.twitter.com/shiv4nsh> | FB <http://www.facebook.com/xeruo> | LinkedIn <https://in.linkedin.com/pub/shivansh-srivastava/3a/204/251>*
