1. V.V. VANNIAPERUMAL COLLEGE FOR WOMEN
(An Autonomous Institution Affiliated to Madurai Kamaraj University)
VIRUDHUNAGAR
(Re-Accredited with “A” Grade (3 rd Cycle) by NAAC)
Online classes for
cloud computing
By
D.Shunmuga Kumari, M.Sc., M.Phil.,
Assistant Professor,
Department of information Technology,
11.08.2020 meet.google.com/zij-ngfi-ipm
V.V. VANNIAPERUMAL COLLEGE FOR WOMEN
(An Autonomous Institution Affiliated to Madurai Kamaraj University)
VIRUDHUNAGAR
(Re-Accredited with “A” Grade (3 rd Cycle) by NAAC)
2. 1.2 Technologies for Network-Based
System
I. Multicore CPUs and Multithreaded Technologies
II. Memory, Storage and Wide-Area Networking
III. Virtual Machines and virtualization middleware
IV. Data Centre virtualization for Cloud computing
3. Technologies for Network-Based System
Efficient and Scalable computing,
Need to explore H/W, S/W, N/W
for distributed computing system design and applications
Focused on distributed operating systems in parallelism approach
A distributed operating system is a software over a collection of independent, networked,
communicating, and physically separate computational nodes. They handle jobs which are serviced by
multiple CPUs.
4. I Multicore CPUs and Multithreaded
Technologies
High performance computing (HPC)
High throughput computing (HTC) systems
Processor speed is measured in MIPS
(Million of instruction per second)
N/W Band width is measured in Mbps
GE Refers to 1 Gbps Ethernet bandwidth
5. I.1 Advanced in CPU Processor
CPU- Multi core architecture with dual, quad, six or more
CMOS based chips due to power limitations
Stands for "Complementary Metal Oxide Semiconductor." It is a technology used to produce integrated circuits. CMOS circuits are
found in several types of electronic components, including microprocessors, batteries, and digital camera image sensors.
their slow speed of operation. Propagation delay time
Instruction level Parallelism- ILP
- include multiple-issue super class architecture, dynamic branch prediction, speculative execution.
-It demands good h/w and compiler support.
-TLP Task/Thread Level Parallelism ,
-DLP Data Level Parallelism - highly explored in graphics programming units(GPU)
7. I.3Multicore CPU, Manycore GPU
Many core GPU with 100 or more of thin cores
IA-32,IA-64 instruction set architecture.
(RISC Processors replaces by)X-86 Processors extend to serve HPC,HTC systems
in high end server processors.
X-86 upgrades will dominate data centres, super computers.
Asymmetric and heterogenous chip multiprocessors in thin GPU.
9. I.4Multi Threading Technology
Thread – 5, Pipelined data paths- 4
4-issue superscalar processors – single thread
Fine grain multithreaded processors- multi threading
Coarse grain multithreaded Processors- executes many thread in few cycles
of instruction switching
Dual core Processors(2-CMP)-execute instruction from different thread
completely.
Simultaneous multithreaded processor(SMT).- simultaneous scheduling of ins
from diff thread in the same cycle.
Blank square – no available ins on time
More blank- lower scheduling efficiency
10. II. Memory storage, Wide Area
Networking
Memory Technology:
Processors get faster, memory capacity is needs in wider space which gives wider gap between
the processors and memory
From 260 MB- 250 GB to 3 TB
Disk Storage Technology:
Rapid growth of flash memory and Solid State Driver(SSD) also impacts the future of HPC
& HTC Systems.
SSD handle 300,000 to 1 million write cycles per block.
Impressive speed up in many applications.
Power consumption, cooling and packaging , Clock rate, voltage applied on chips needs
to redefine
System Area interconnection:
11. II. Memory storage, Wide Area
Networking
System Area interconnection:
LAN – small clusters are interconnected – client to server
SAN – Server Area Network – connects server to network storage.
NAS – Network Attached Storage –
connects client host to disk arrays
12. III. Virtual Machines and Virtualizations
Virtual Machines:
Offers novel solutions to underutilized resources , applications inflexibility, s/w manageability ,and
security concerns in existing machines.
Cloud resources rely on virtualizations of processors, memory and I/O facilities in dynamically.
13. III. Virtual Machines and Virtualizations
VM Primitive operations:
a) Multiplexing: mux bet h/w machine
b) Suspension : Stored in stable storage
c) Provision : can be resumed
d) Life migrations migrated from
one h/w platform to another
VM operations enables a VM to be provisioned to any available h/w Platform. They also enable
flexibility in porting distributed application executions. It approaches will significantly enhance
the utilization of server resources.
14. IV. Data Center virtualizations for cloud
computing
Large data centers are built with 1000 of servers. (Small data center – 100 of
servers)
Low cost design philosophy
high-band width networks may not fit the economics of cloud computing.
s/w layer handles n/w traffic balancing, fault tolerance and expandability
Convergence of technologies
H/w Virtualizations and multi core chips
Utility and grid computing
Web 2.0
Autonomic computing and data center automation