how to implement a poller pool on server? - SNMP

This is a discussion on how to implement a poller pool on server? - SNMP ; Hi, I am pretty new to this but here is the problem and I may be posting to the wrong group, but I'd give it a try since SNMP people know a lot about network and server programming I am ...

+ Reply to Thread
Results 1 to 2 of 2

Thread: how to implement a poller pool on server?

  1. how to implement a poller pool on server?

    Hi,

    I am pretty new to this but here is the problem and I may be posting
    to the wrong group, but I'd give it a try since SNMP people know a lot
    about network and server programming

    I am creating a web application used for hospital, the portal displays
    real time information form medical device (such as blood pressure,
    whatever). There will be say, tens or hundreds of users using the
    portal, each requests information from different individual device.

    The devices will be hooked to a server (or a server farm, depends on
    the scale). The server needs to serve all the requests from all the
    browsers.

    My impression is that I can use a pool of poller in server (size
    depends on reasonable number of concurrent users, and maybe able to
    dynamiclly increase pool size, but that is not my question). And I
    plan to use a web service to expose the data request/serve function.

    The questions are:
    1. how do I implement of this poller pool?
    2. how do I associate the poller with web service? (like, how can I
    use a getDeviceData(deviceID) call to make a poller to poll device
    with ID, and get the data returned by poller?

    If the poller always polls a fixed device, then it is easier, but the
    problem is that poller is idle until request comes in with a certain
    device ID so it is very dynamic.

    Any suggestion is appreciated!

    Sally

  2. Re: how to implement a poller pool on server?

    On Oct 24, 12:12 am, an.sa...@gmail.com wrote:
    > Hi,
    >
    > I am pretty new to this but here is the problem and I may be posting
    > to the wrong group, but I'd give it a try since SNMP people know a lot
    > about network and server programming
    >
    > I am creating a web application used for hospital, the portal displays
    > real time information form medical device (such as blood pressure,
    > whatever). There will be say, tens or hundreds of users using the
    > portal, each requests information from different individual device.
    >
    > The devices will be hooked to a server (or a server farm, depends on
    > the scale). The server needs to serve all the requests from all the
    > browsers.
    >
    > My impression is that I can use a pool of poller in server (size
    > depends on reasonable number of concurrent users, and maybe able to
    > dynamiclly increase pool size, but that is not my question). And I
    > plan to use a web service to expose the data request/serve function.
    >
    > The questions are:
    > 1. how do I implement of this poller pool?
    > 2. how do I associate the poller with web service? (like, how can I
    > use a getDeviceData(deviceID) call to make a poller to poll device
    > with ID, and get the data returned by poller?
    >
    > If the poller always polls a fixed device, then it is easier, but the
    > problem is that poller is idle until request comes in with a certain
    > device ID so it is very dynamic.
    >
    > Any suggestion is appreciated!
    >
    > Sally



    Hi Sally,

    You can do two ways
    method 1) you create an application which periodically polls all the
    devices and keeps the collected data in database. A web server have
    some mechanism to retrieve data from the database on a request from
    web browser.
    you have choice of languages for application
    - C -- netsnmp
    - perl has
    - java
    - C++
    You have choice of web technology to access DB
    - webserver ( say apache) CGI/perl or CGI/others with DB modules
    - java servlets with jdbc
    - others PHP,python and more

    method 2) browser(reqX)->webserver->web service-->request goes to
    device
    and comes back.
    web technology can be same as in method 1
    each new request is generated at SNMP layer for each
    request from browser. No pooling required unless you want to
    restrict max requests at any particular time. You can have a max limit
    check with
    a semaphore in unix.
    However to limit requests per device kind of case,
    java servlet could do it easier than others as it requirs more book
    keeping between different requests.

    method 1 is more useful.
    method 2 is simple (if you do not have per device req limit)

    - Surinder

+ Reply to Thread