Considering the Map-Reduce paradigm and the Hadoop platform, jobs often run on inexpensive hardware where faults may be frequent.

How is Hadoop able to carry out Map-Reduce jobs to completion even in the presence of problems?

(select the option that best fits the question).

One of the core contributions of Spark is the proposal of the Resilient Distributed Datasets (RDDs). 

What are the core features in RDDs that provide Spark with fault-tolerance when executing jobs and workers fail?

(select all the correct options).

Method inlining is a key and frequently employed optimization to achieve high-performance emulation in Java.

Taking into consideration how method inlining can be implemented and when it should be activated, identify the correct sentence(s).

(select all the correct options).

In the context of binary translation, consider a scenario where the execution is dominated by the repetitive emulation (inside a loop) of the following sequence of translated blocks of code: A, B, C.

- Block A is 5 instructions long, 

- Block B is 10 instructions long, 

- Block C is 15 instructions long, 

- The Emulation Manager requires 5 instructions to execute.

- Each jump instruction takes (naturally) 1 instruction to execute.

What is the most appropriate optimization to employ in this scenario?

What is the average speedup in emulation that could be be obtained by applying such an optimization?

Justify your answers and provide brief reasoning for calculations with the best information available.

(No penalty for wrong answer).

The interaction between each AWS instance with other instances and with the outside world must be monitored.

What is the key mechanism in AWS that allows developers to define how communication between EC2 instances can be configured to be restricted or allowed?

(select the option that best fits the question).

In the context of a MapReduce deployment in Hadoop, when there are several users and several MapReduce jobs to execute in the same shared infrastructure, there needs to be a scheduler deciding when to execute jobs and what resources to allocate to each one running.

From the following sentences, what is the one that best reflects how job scheduling in MapReduce takes place?

(select the option that best fits the question).

Consider a scenario where a city is hosting an international music festival event expected to have several hundreds of thousands of people attending.

The city commissioned a public infrastructure authority to set up the basic infrastructure for lighting (i.e. electricity cables for lamp posts) and the points-of-sale (i.e., both electricity and Internet/data cables for vending machines and cash registers at bars). 

The area where the event is being held is divided as a grid of 10 x 10 sub-areas called sectors, and named from A0 to J9. 

Each sector is a 1000m x 1000m square. Each sector has one electricity access point and one data access point at its geometric center, i.e., 500m x 500m point inside the sector.  Each location has global positive x and y coordinates, with the top-left corner of sector A0 being (0, 0) and the bottom-right corner of J9 being (9999,9999).

The infrastructure is going to be laid out in parallel across the 100 sectors. Each sector has a dense grid of pathways (for bicycles, pedestrians, etc.), along with trenches where cables can be placed/installed/laid out.  

Therefore, the infrastructure authority wants to calculate the extension of electricity cables and data cables that are needed, in each sector, to properly connect all the lamp posts and points-of-sale directly to the relevant access points in that sector (e.g., sector A1 needs 4000m of electricity cable and 400m of data cable).

In general, the authority needs to iterate over the multiple locations proposed to place lamp posts and points-of-sale, and determine based on their global coordinates:

- what cables to connect, where to connect them to, and therefore calculate the extent of electricity and/or data cable needed, so that totals for each specific sector can be obtained.

The input consists of one text file.

In the file locations.txt, each line stores one intended location for a lamp (post-type=LAMP) or a point-of-sale (post-type=SALE):     

<postID>, <post-type>, <x-coord>, <y-coord> 

Assume the standard signature for the mapper: List<k2,v2> map(int, string) and reducer:  List<k3, v3> reduce(k2, List<v2>) functions, where map has the default MapReduce input (cursor/line number and text line content) and k2, v2, k3 and v3 can all be replaced with types of your choice.  

Write the pseudo​code for a MapReduce application, with a single ​map ​and a single ​reduce ​method.

The final output of the reduce function should produce, if required:

- a single record for each sector, identifying it, and with two data fields, stating in meters the total extent of electricity cables and of data cables required to properly connect all the locations in that sector.

(if you need, you can briefly state any assumptions at the end of the answer after the pseudocode.)

(No penalty for wrong answer.)

Note: MapReduce is not natural language. Pseudocode needs not be Java but it must include clear operating steps well identified, it should not be a paragraph or set of paragraphs rephrasing the question. But do express your reasoning, nonetheless.

Regarding the cloud storage Dynamo from Amazon, identify exactly a specific key design decision.

(select the option that best fits the question).

In a cloud data center facility where the total energy consumed is 5000 kW and the energy delivered to computing equipment is 4000 kW, what is the DCiE (Data Center Infrastructure Efficiency) of this facility?

(select the option that best fits the question.)

Consider an Infrastructure-as-a-Service deployment, in order to seamless extend computing resources for a small start-up company, in a real-life scenario employing replicated m front-ends and n replicated back-ends. 

When redirecting requests from many clients to the web application among a number of virtual servers, what situation can happen in terms of scale limitations and/or failures?

(select the option that best fits the question).