Collect statistics and analysis of google trace - open to bidding
$30-250 USD
Πληρώθηκε κατά την παράδοση
Hello
I need someone to program in ipython and collect statistics from google trace
You need to download the trace
[url removed, login to view]
Delivers in 3 days (Thureday 4 August) which include:
ipython program using anaconda with clear comments and be as simple as possible
a report about the statistics and the analysis with tables and figures and details
Statistics include:
1. Degree of Parallelization (DOP)
We begin by studying the relationship between the degree of parallelism of a job and its final status. In Google’s cluster, a job consists of one or more tasks that typically execute the same binary with the same resource requirements and scheduling constraints (e.g. priority, scheduling-class, etc). Applications that need to run different types of tasks will usually execute them as separate jobs. For example, MapReduce applications would execute masters and workers as separate jobs. Generally, multi-task jobs are meant to have their tasks run simultaneously, where a single task can be running on a single machine at any point in time. A configuration parameter is available where a user can indicate if tasks must execute on different physical machines.
2. Requested Resources: We now study the relationship between the amount of resources requested by tasks in a job the final status of the job. In Google, tasks are submitted with values for requested CPU, memory, or disk space, where these values represent the maximum amount of resources a task is allowed to consume on a machine. However, tasks are sometimes permitted to use more than what they requested if resources are available; e.g. tasks may use free CPU cycles on a machine .
The following materials in the attachment can help you in collect statistics (they have the same statistics with results)
Some helpful references:
C. Reiss, J. Wilkes, and J. L. Hellerstein. Google cluster-usage traces: format + schema, 2011. [url removed, login to view]
J. Wilkes, “More Google cluster data,” Google research blog, Nov. 2011, Posted at [url removed, login to view] less
Ταυτότητα Εργασίας: #11182745
Σχετικά με την εργασία
1 freelancer κάνει προσφορά κατά μέσο όρο $431 για αυτή τη δουλειά
Hello Let's explore the requirement and kindly let us know if you would like us to share our skills & experiences with previous development. Thanks & Regards Moumita