In the Microkernel architecture of VMs:
Drivers are part of the Guest VMs
Drivers are part of the Hypervisor
Drivers are part of a specific service VM
Drivers are part of the Guest OS
Which statement about Paravirtualization is correct?
It is the same of Kernel-level Virtualization
Guest OS and VMM are independent each other
It cannot be used with traditional Operating Systems
Hooks are not required
Which is the goal of the "machine-learning software as a service"?
Provide networking paradigms for machine learning
Provide ready-to-use machine learning applications
Provide platform and toolboxes for machine learning
Provide hardware infrastructures for machine learning
An emulator:
emulates only the ISA but not the ABI
emulates only the ABI and not the ISA
reproduces the machine by providing not only ABI but ISA (user/ system)
supports the Java Runtime Environment
Does an emulator read, interpret and execute all the bytes included in the memory of system it is going to reproduce?
The program in warehouse-scale computing may consist of tens or more individual programs:
Data centers are buildings where multiple servers and communication units are co-located
What is a computing infrastructure?
Technological infrastructure that provides hardware and software for computation to other systems and services
Technological infrastructure that provides applications to other systems and services
Tower servers consumes less space than rack and blade:
Blade servers require expensive configurations:
TOR switch refers to:
Tower switch
Top-of-rack switch
Racks shelves may contain switches:
What is a Network Attached Storage (NAS)?
a storage system directly attached to a server or workstation
a computer connected to a network that provides only file-based data storage services to other devices
a remote storage unit connected to PC using a specific networking technology
a computer connected to a network that provides computation to other devices
Downtime is a period of time, or a percentage of a time span, when a system is unavailable or offline __________
none of the other
because of routine maintenance
because of system crashes
regardless of the cause
In a three-layer network architecture of a Datacenter, which is the layer that is typically associated to the TOR switch?
Access
Core
Aggregation
None of the others
Which definition is not encompassed by Dependability:
Reversibility: ability to reverse a broken service
Reliability: continuity of correct service
Maintainability: reparation to restore correct service
Availability: readiness for correct service
In the context of SSDs, Blocks are the smallest unit that can be read/written
Paravirtualization is a particular type of kernel-level Virtualization that does not rely on a host OS
The structure (or topology) of a reliability block diagram is always the same as the structure (or topology) of the modeled system
A system virtual machine provides a complete system environment that can support an operating system (potentially with many user processes)
Which sentence about Tier Level 4 of datacenter is not correct?
Fault-tolerant site infrastructure with availability of 99.995%
Meets or exceeds all Tier 3 requirements
Single non-redundant distribution path serving the IT equipment
All cooling equipment is independently dual-powered
What is a hybrid cloud computing system?
a cloud used for a single organization
a cloud whose energy is produced both by diesel and batteries
a cloud shared by several organization
a cloud composed by private and public cloud systems
Paravirtualization cannot be used with unmodified OS
Servers in data centers are like regular personal-computers with the same form factor
Warehouse-scale computers run a large number of very small applications
An Edge Computing system provides higher performance (in terms of memory and computation) than Internet-of-Things devices:
Which of the following is not a feature of a fat tree topology in a data center network architecture?
It has multiple connections to the core
It has a recursive organization
It has a three-tier model
In the context of SSDs, one of the following activities is NOT managed by the Flash Translation Layer:
Defragmentation
Address Translation
Garbage Collection
Data Allocation
A System VM:
Provides the interface to system ISA and user ISA
Provides the interface to system ABI and the user ISA
Provides the interface to system ABI and the user ABI
Provides the interface to system calls and user ABI
Consider an array of N disks. Select the right order of the RAID levels in terms of effective data storage in ascending order, i.e. for the lower to the higher data storage availability.
RAID 1+0, RAID 6, RAID 5
RAID 6, RAID 5, RAID 1+0
RAID 1+0, RAID 5, RAID 6
Warehouse-scale computers belong to a single organization, use a relatively homogeneous hardware and system software platform, and share a common systems management layer
All faults occurring in a system cause a failure of the system
In the context of SSDs, Pages are the smallest unit that can be erased
Graphical Processing Units support data-parallel computation
A deep neural-network cannot be trained on multiple Graphical Processing Units
A System Virtual Machine can provide its functionality:
Only on the hardware
Only on top of an Operating System
Either working directly on the hardware or running on an Operating System
Hybrid Clouds are systems where one or more public cloud services are combined with private data center resources into a single virtually managed environment.
Graphical Processing Units within a tray are connected using high-bandwidth interconnects, such as NVlink:
The need for a constant Internet connection is one of the advantages of data centers
In-rack cooling refers to cooling equipment where the cooling coils are adjacent to the rack.
In datacenter storage context, DAS means Distributed Application Storage:
Which is the difference between Guest and Host?
Guest: the underlying platform supporting the environment/system; Host: the software that runs in the VM environment
Guest: the software that runs in the VM environment; Host: the underlying platform supporting the environment/system
Guest: the underlying platform supporting the environment/system; Host: the virtual machine manager supporting the execution of virtual machines
Guest: the underlying platform running in the virtual machine manager; Host: the software that runs in the VM environment
To increase reliability, which of the following actions is not correct?
Use multiple redundant components
Reduce MTTR to a minimum
Have spare components at disposal
Use elements with higher MTTF
Warehouse-scale computers can be designed to operate on multiple data centers
Rotational delay is the time to rotate the read head to reach the desired track
A system having very small reliability may have large availability
In the context of data centers, scalability is the ability for the infrastructure __________ without cost, efficiency, and reliability being compromised. Select the correct item to be included in the previous sentence.
to be enlarged or to handle an increment of the input requests
to run in a larger building
none of the others
to improve the cooling strategy
Which is the most appropriate definition for "fault"?
Any library inconsistency in the Operating System
Any defect in the silicon caused by dust particles during manufacturing
Any inappropriate usage of the system
Any event interfering with the nominal behavior of the system
One of these levels is NOT part of a typical three-layers network architecture of a data-center?
Cloud
In the context of rack servers, 1U is 44.45cm
In a reliability block diagram, adding components in series decrease the system reliability with respect to the reliability of the single component
What are the benefits of data center virtualization?
All of the other answers listed are correct.
Consolidation and ease of manageability.
Cost savings in power and cooling of the data center.
Flexibility and agility.
When dependability-related requirements must be considered?
Only during requirements definition
When starting testing/verification activities
When a prototype of the system is available
In every phase of the design flow
The program in warehouse-scale computing is a service provided by internet:
Traditional data centers can host hardware and software for multiple organizational units or different companies.
Which is the configuration characterizing the three-layers network architecture of a data-center?
Sensors – Aggregation – Cloud
Access - Aggregation – Cloud
Access – Aggregation – Core
Access – Fog – Core
Tower servers have a cable management that is more complex than Rack ones:
Gmail is an example of:
Software as a Service
Hardware as a service
Platform as a Service
Infrastructure as a service
Consider an array of N disks. Select the right order of the RAID levels in terms of effective data storage in descending order, i.e. for the higher to the lower data storage availability.
RAID 0, RAID 4, RAID 6
RAID 6, RAID 0, RAID 4
RAID 4, RAID 6, RAID 0
Data centers are well suited for real-time decisions in autonomous car context
After a fault activation the system is always expected to fail
The reliability of a Triple Modular Redundancy (TMR) system is always greater than the reliability of the single copy of the system
What is a Data Center?
A building or portion of a building whose primary function is to research new data.
A building or a portion of a building whose primary function is to house a computer room and its related support areas.
A room in a building dedicated to monitoring data and information.
Seek delay is the time needed by the disk to place the desired sector under the read head
What is the Hardware-level virtualization?
The virtualization layer is placed between the operating system and the applications
The virtualization layer is above the applications
The virtualization layer is placed between the hardware and the operating system
The virtualization layer is placed below the hardware
Which aspect is a drawback of data center computing platform?
Power consumption
Device independence
Universal access
Unlimited storage capacity
Racks shelves are used to store servers and other IT equipments:
In warehouse-scale computers the software running on these systems executes on clusters of hundreds to thousands of individual servers
In a data center for AI, servers includes both general purpose CPUs and accelerators (such as GPUs, TPUs and FPGAs):
NVlink is a low-bandwidth connection among Graphical Processing Units:
A process VM:
Provides the interface to system calls and user ISA
One of these configurations is not considered in Data-center network architectures. Which one?
D-Cell
Three layers
Fat-tree
Direct Connections
What is the average latency for an Hard Disk Drive?
Time to wait for the sector
Time to wait for the data transfer
Time to wait for the head movement
Time to wait for the cylinder
In the context of SSDs, Pages are the smallest unit that can be read/written
In-row cooling adds an air-to-water heat exchanger at the back of a rack so the hot air exiting the servers immediately flows over coils cooled by water.
Select the wrong definition. A Virtual Machine:
may appear as having different resources than the physical machine
may result in a different level of performance with respect to the physical machine
provides identical software behavior
Tower servers can be cooled easily:
What is the TBW in a Solid State Disk?
Total amount of data that must be written into an SSD before the completion of the space
Total amount of data that can be written into an SSD before it is likely to fail
Total amount of data that can be read from an SSD
Total amount of data that can be written into an SSD
The following sentence: "The probability that the component does not fail in the interval 0 ... t, knowing that at the instant t = 0 the component was working correctly" defines:
Dependability
Reliability
Availability
None of the other
Consider an array of N disks. Select the right order of the RAID levels in terms of MTTFRAID in descending order, i.e. for the higher to the lower MTTF. Consider a number of disks N that is large enough to satisfy the minimum requirements for all the RAID levels.
RAID 5, RAID 1+0, RAID 0+1
RAID 1+0, RAID 0+1, RAID 5
RAID 0+1, RAID 5, RAID 1+0
Racks servers have a cable management that is more complex than Tower ones:
Traditional data centers typically host a large number of relatively small- or medium-sized applications
In-row cooling works like in-rack cooling except the cooling coils are not in the rack, but adjacent to the rack.
Rack, blade and towers are form factor of servers in data centers
Warehouse-scale computers comprise important components related to power delivery, cooling, and building infrastructure
What is a Graphical Processing Unit (GPU)?
A specialized hardware meant to provide storage in a data center
A specialized hardware meant to support data-parallel computation
An Internet-of-Things device meant to operate in the field
A specialized hardware meant to provide networking in a data center
Rack servers need additional cooling systems due to their high overall component density
Is the Java Runtime Environment an example of system virtual machine?
Which is the main characteristic of the D-Cell topology for data center network architectures?
The network is organized in a distributed way
The network comprises edge, aggregation and core layers
the network is organized in a recursive way
the network is organized in a hierarchical way
