Mot otiv ivat ation, ion, Logis Logistics ics, , and and Int - - PDF document
CMPE 252A: Computer Networks J.J. Garcia-Luna-Aceves Office: E2 317 jj@cse.ucsc.edu phone: 4153 http://users.soe.ucsc.edu/~jj/CLASSES/CMPE252A-FALL2016/ 1 CMPE 252A: Computer Networks SET 1 Mot otiv ivat ation, ion, Logis Logistics
1
http://users.soe.ucsc.edu/~jj/CLASSES/CMPE252A-FALL2016/
2
3
JJ
4
My research: Computer communication Internet Wireless networks Distributed algorithms Mobile and pervasive computing Cyber-physical systems
JJ
JJ
https://users.soe.ucsc.edu/~jj/
To become familiar with the field of
Network architectures, protocols, and
To learn basic tools for the verification and
To learn about a few specific protocols To get practice in the art of reading
Each class will cover a set of papers and book chapters Readings are assigned ahead of time and you must
As part of homeworks, you must submit your answers to
I will call on students to answer questions based on
Pace of class is fast and assumes you are doing your
I may pile more slides and papers than can be covered
No curves Three components:
Two midterms account for 40%:
Homework assignments account for 25% Project accounts for 35%:
Content: 50%; Report: 30%; Presentation/Demo: 20%
Dates for exams and homeworks will be
No textbook On-line books in class page Papers and class notes are the sources of
Will be posted before and after lectures.
You must go over the class notes and
Use class notes as your study guide.
9
Focus your reading
What problem is the paper solving? What is the main idea and what do you think of it? Can you verify [and hence trust] the results? How well is the paper written?
Layered approach to reading research
Do a quick read in 5-10 minutes Read again to understand the insight, contribution and
intuition of results; ignore proofs and analysis
Go back and study details Deconstruct paper; question all assumptions
10
Take your time and ask questions:
Do you understand the proofs? Any errors? Question the assumptions being made Any prior work missing? Is the message of the paper substantiated by
How else would you present the information? Do you have different/better ideas?
11
12
How we do business
E-commerce, advertising, cloud computing, connected robots
How we learn and are informed
Google, Wikipedia, on-line courses, on-line news generators
How we do science and health
Google, access to big data, tele-presence
How we socialize and entertain
E-mail, IM, Facebook friends, virtual worlds, on-line casinos
How we think about law
Interstate commerce? National boundaries? Wikileaks?
How we govern
E-voting, e-government, censorship, wiretapping
How we fight
Cyber-attacks, including nation-state attacks, cyber-physical wars
Infrastructure-centric computing
(e.g., mainframes, time-sharing, IBM, Sun Microsystems)
Infrastructure-centric networks
(e.g., telephone network, cellular networks, ARPANET, emergence of Internet; think AT&T, BBN, Cisco, Nokia)
Very expensive links, storage and processing, independent
vertical systems
Focus on establishing the infrastructure Research: How do we use links and processors as
efficiently as possible?
14
Present:
Information-centric computing
(e.g., the web, personal computing devices; Apple, Google)
Information-centric networking and the cloud
(e.g., Content-Delivery Networks, caching systems, cloud computing, firewalls; think Akamai and Oracle)
Programmable networking
(SDNs, Openflow, Nicira)
Affordable links, very cheap storage and processing, network and
service integration (e.g., VoIP calls), the cloud, Openflow
Focus on enabling services and information everywhere Research: How do we find and replicate information efficiently?
How do we process information efficiently and securely at remote servers? How do we represent the network at controllers?
15
Unit sales of industrial robots
Source: Swarming MAVs Form Insta-Network http://www.botjunkie.com/index.php?s=landroid
Any map from (x2,y2) at time t2 Urgent for all workers in region (x,y) at time t0 to t2 Any drone near (x2,y2) at time t2 21
7.2 6.8 7.6
Rapid Adoption Rate of Digital Infrastructure: 5X Faster Than Electricity and Telephony
50 2010 2015 2020 40 30 20 10
BILLIONS OF DEVICES 25 12.5 Inf nflect lection ion Point
TIMELINE Source: Cisco IBSG, 2011
World Population
Around 2008, number of connected things > world population
Sensors connected to and
discoverable on the Web
Sensors have position &
generate observations
Sensor descriptions available Services to task and access
sensors
Local, regional, national
scalability
Enabling the Enterprise Webcam Environmental Monitor Industrial Process Monitor Stored Sensor Data Traffic Monitoring Satellite-borne Imaging Device Airborne Imaging Device Health Monitor Strain Gauge Temp Sensor Automobile As Sensor Probe Shopping
People love sharing!
We have gone from letters to non-electric telegraphs
to electric telegraphs to telephone calls to IM and sharing files mostly at remote servers to twitting “whatever” … We are at the “twitting point” of sharing Device revolution!
In < 50 years we have gone from “the SRI van”
hosting a limited packet radio to PDAs that are many
Why not share everything,
Revolution in devices:
Sensors, processing and storage in the ambient are
becoming commodities
24 billion connected devices in the world by 2020 Over half of them will be such non-mobile devices as
household appliances
Connected devices will be a US$1.2 trillion market.
Revolution in services:
People expect services on the move within social contexts
Yet, autonomic, scale-free wireless
Big data:
A commercial jet generates 10 TB for every 30
Genomics data: More than 10 times Youtube
Billions of sensors 24/7 augmented reality; data rate per person?
Speed of light and large bandwidth
White spectrum Long distances Need to store/process big data in the network
User mobility and multi-homing
26 billion mobile devices expected by 2020 The “typical” end host is now nomadic Cloud services are virtualized/replicated/
V(X)LANs are equally mobile/dynamic
Security, privacy, integrity
Data, services, and things Wasting the network with useless data
Seamless, autonomic operation
The Internet has to be everywhere
Simple interaction with services, data, things
Customer-driven apps, services, and
If user has to see it to use it, we are
Gartner “Hype Cycle”, July 2014
Expectations = Funding Opportunities and Jobs
Human augmentation Software-defined anything Connected home Digital security Internet of Things
31
Social Networks Information Networks
Networks Networks on switches and chips, Nano-networks
32
Physical connectivity, node mobility, link quality Infrastructure plane Information plane social plane User preferences, group mission and tasks, Information replication, social group descriptions
33
analytical models fundamental limits
Upper-bound Lower-bound
simulations
Protocols and Architectures for Computer Networks
PHYSICAL LINK NETWORK TRANSPORT APPLICATION neighborhood discovery transmission scheduling Antennas, radios interconnection collaborative applications… end-to-end transport protocols… Routing Congestion packet forwarding
logic
Communication networks, social networks, smart grid, game networks, IoT, etc.
35
Nicolaus Copernicus (1473-1543): Polish astronomer and
mathematician who was a proponent of the view of an Earth in daily motion about its axis and in yearly motion around a stationary sun.
36
37
RTP DNS UDP TCP IP
Ethernet
ATM
WiFi
Optical Fiber
FTP
CURRENT INTERNETWORKING ARCHITECTURE
HTTP UWB routing
telnet 38
“Copernican revolution” in content:* Exploding content-driven market! Google & Yahoo! vs IBM & BUNCHS Internet is access to content and VoIP
RTP DNS UDP TCP IP
Ethernet
ATM
WiFi
Optical Fiber
FTP
CURRENT INTERNETWORKING ARCHITECTURE
HTTP UWB routing
telnet
Good: Connectivity to any host through any transmission medium has been attained.
“Gutenbergian revolution” in resources:♣ Exponential growth in number of end user (wireless) devices, and their processing and storage capacities.
* M. Gotta et al., “The Copernican Revolution in Content,” Burton Group, 2006
Bad: Focus has been only on connectivity! No storage or computing. Users and content were afterthoughts.
♣ JJ, this class
39
Focus: The best match among content, services, resources, and customers
UDP TCP
IP
USER AND CONTEXT CENTRICITY
Ethernet ATM WiFi Optical Fiber UWB
content applications
storage
computing
networks DNS routing
RTP
FTP HTTP
telnet
“Invisible Internet” with very simple protocols and mechanisms Simple, small, cost effective Location-independent, policy-based services. Social, secure, context-aware Focus on utility to end-user needs & predicate-based interaction. Easy, state what not how
40
41
A communication network is a set of nodes connected
by links and able to communicate with one another.
A computer network is a communication network
in which nodes are computers.
The purpose of the network is to serve users, which can be
humans or processes.
Network links can be point-to-point or multipoint and
implemented with several transmission media.
Information exchanged can be represented in multiple
media (audio, text, video, images, etc.)
Services provided to users can vary widely. 42
We will use the Internet and standard
The Internet has computer hardware, software,
Communication protocols implemented in
Protocols specify how processes in different
Distributed algorithms are the essence of what we
43
A set of rules governing the interaction of
A protocol has five parts:
The service it provides. The assumptions about the environment where it
executes, including the services it enjoys.
The vocabulary of messages used to implement it The format of each message in the vocabulary. The procedure rules (algorithms) guarding the
consistency of message exchanges and the integrity of the service provided.
44
What is a good protocol design?
Judging by their survival, Ethernet and IP are good;
token ring protocols are not very good
What are good and bad aspects in a protocol?
TCP adapts to congestion, but it inherently assumes that
the Internet sends packets in order.
Use representative protocols to go over these
Discuss new directions in computer
45
We will try to focus on the
Our principles are:
The description of a protocol has no ambiguity. A protocol does what it is supposed to do, all the time. A protocol does not leave any communicating party
waiting forever for something to happen.
A protocol makes efficient use of available resources. A protocol enables the use of resources fairly or
according to a predefined contract.
As with most engineering topics, simplicity is important.
46
Protocol specification: The description of the
Safety: A protocol does what it is supposed to do,
Liveness: A protocol does not leave any
Efficiency: A protocol makes efficient use of
Fairness: Fair or contractual use of resources Simplicity is desirable, but hard to attain…
47
Very few researchers have paid any attention to
There is no equivalent of Shannon’s information
Be alert!
Challenge conventional wisdom: A protocol is not great
just because a “Big-Name dude” proposed it.
Trust mathematical proof, not qualitative arguments Consider the limitations of any analytical or discrete-
event simulation model, and the limitations of any testbed or deployment
48