Transcript slides - Temple University

Principles of Data Management
Syllabus & Intro
Welcome!

Course website:
 http://cis.temple.edu/~edragut/CIS5516fall2014/teaching.htm
Important Pre-Req
 Text Book(s)
 Workload
 Intended Schedule
 Projects

 Grading

Papers
What Is a DBMS?
A very large, integrated collection of data.
 Models real-world enterprise.




Entities (e.g., students, courses)
Relationships (e.g., Madonna is taking CS564)
A Database Management System (DBMS) is a
software package designed to store and
manage data.
Files vs. DBMS
Application must stage large datasets
between main memory and secondary
storage (e.g., buffering, page-oriented access,
32-bit addressing, etc.)
 Special code for different queries
 Must protect data from inconsistency due to
multiple concurrent users
 Crash recovery
 Security and access control

Why Use a DBMS?
Data independence and efficient access.
 Reduced application development time.
 Data integrity and security.
 Uniform data administration.
 Concurrent access, recovery from crashes.

Why Study Databases??

Shift from computation to information



at the “low end”: scramble to webspace (a mess!)
at the “high end”: scientific applications
Datasets increasing in diversity and volume.



?
Digital libraries, interactive video, Human
Genome project, EOS project
... need for DBMS exploding
DBMS encompasses most of CS

OS, languages, theory, AI, multimedia, logic
A Brief DB History

Early 1970s
 Many database systems
 Incompatible, exposing many implementation details

Then Ted Codd came along




Relational model
Structured Query Language (SQL)
Implementation differences became irrelevant
A few major DB systems dominated the market
Then Web 2.0 & 3.0, Big Data Happen
 What
do you think happen?
 Semi-structured data happen.
•A lot of it and in many forms…
Some Facts about Web x.0 and Big
Data



Twitter: 255 million monthly active users and 500
million Tweets are sent per day,
Facebook: over 1 billion monthly users and faces 3
million message per 20 minute
Instagram: 200 Million Monthly Active Users and 1.6
Billion Likes and 60 Million Photos shared every day
Database Systems Landscape Nowadays
Somebody, Please, Bring Some
Order to This Madness – Cont’d
Somebody, Please, Bring Some
Order to This Madness – Cont’d

NoSQL Databases
Somebody, Please, Bring Some
Order to This Madness




Different Interfaces
Different hardware
support
Different application
support
Lack of Uniformity
Source: http://www.infoq.com/articles/State-of-NoSQL
Database Evolution Timeline
Additional Resources
Tutorial by C. Mohan, An In-Depth Look at
Modern Database Systems
 https://docs.google.com/file/d/0B7lNUaak
0bK1encwYnBVUWZSWjA/edit

Relational Data

Tables or Relations
Relational Database: Schemas
Relational Database: Query
Language
 SQL
- Structured Query Language
 a declarative language designed for
managing data held in a relational database
management system
• Tell what you want and from where
• Do not tell: how to get the data
Key-Value Store



Implemented as an associative array, map, symbol
table, or dictionary abstract data type composed of
a collection of (key, value) pairs such that each
possible key appears at most once in the collection.
A simple put/get interface
Great properties: scalability, availability, reliability
Key-Value Store Usage Scenarios

Increasingly popular within data centers and in P2P
Data center
amazon.com
LinkedIn
Facebook
Dynamo
Voldemort
Cassandra
P2P
Vuze
uTorrent
Vuze DHT
uTorrent DHT
Row Store and Column Store
Source: Column-Oriented Database Systems, VLDB 2009. Tutorial; S.
Harizopoulos, D. Abadi, P. Boncz



In row store data are stored in the disk tuple by tuple.
Where in column store data are stored in the disk column by
column.
Column-stores are more I/O efficient for read-only queries as
they read, only those attributes which are accessed by a query.
Row Store and Column Store
Row Store
Column Store
(+) Easy to add/modify a
record
(+) Only need to read in
relevant data
(-) Might read in
unnecessary data
(-) Tuple writes require
multiple accesses

So column stores are suitable for read-mostly,
read-intensive, large data repositories
Graph Databases
Biological
Network
Ecological
Network
Social
Network
Chemical
Network
Program Flow
Web Graph
Database Management Systems 3ed, R. Ramakrishnan and J. Gehrke
23
Graph Databases: Query

Find all the restaurants my friends (in Facebook) like
So, What Does CS Curricula Cover?

Undergrad Database Courses
 Introduction to Relational Databases

Grad Database Courses
 Advanced Relational Databases in IS&T
 Principles of Data Management in CS
• Still relational DBMS.

How about the rest of database system types?
 Good question…

Where is most of the research activity going on nowadays?
 Good question again…
So, Why Study Relational DBs?

Jack Clark, The Register, 30 August 2013: “The tech world is turning
back toward SQL, bringing to a close a possibly misspent half-decade in
which startups courted developers with promises of infinite scalability
and the finest imitation-Google tools available, and companies found
themselves exposed to unstable data and poor guarantees.”

Google Spanner paper, October 2012: “We believe it is better to have
application programmers deal with performance problems due to
overuse of transactions as bottlenecks arise, rather than always coding
around the lack of transactions.”

Sean Doherty in Wired, September 2013: “But don’t become
unnecessarily distracted by the shiny, new-fangled, NoSQL red buttons
just yet. Relational databases may not be hot or sexy but for your
important data there is no substitute.”
And, The Key Reason of All

Gartner estimates RDBMS market at $26B with
about 9% annual growth, whereas Market
Research Media Ltd expects NoSQL market to
be at $3.5B by 2018.
 Source: C Mohan’s tutorial
Data Models
A data model is a collection of concepts for
describing data.
 A schema is a description of a particular collection
of data, using the a given data model.
 The relational model of data is the most widely
used model today.



Main concept: relation, basically a table with rows and
columns.
Every relation has a schema, which describes the
columns, or fields.
Levels of Abstraction

Many views, single
View 1 View 2 View 3
conceptual (logical) schema
and physical schema.
Conceptual Schema



Views describe how users
see the data.
Conceptual schema defines
logical structure
Physical schema describes
the files and indexes used.
Physical Schema
 Schemas are defined using DDL; data is modified/queried using DML.
Database Management Systems 3ed, R. Ramakrishnan and J. Gehrke
29
Example: University Database

Conceptual schema:




Physical schema:



Students(sid: string, name: string, login: string,
age: integer, gpa:real)
Courses(cid: string, cname:string, credits:integer)
Enrolled(sid:string, cid:string, grade:string)
Relations stored as unordered files.
Index on first column of Students.
External Schema (View):

Course_info(cid:string,enrollment:integer)
Data Independence *
Applications insulated from how data is
structured and stored.
 Logical data independence: Protection from
changes in logical structure of data.
 Physical data independence: Protection from
changes in physical structure of data.

 One of the most important benefits of using a DBMS!
Concurrency Control

Concurrent execution of user programs
is essential for good DBMS performance.

Because disk accesses are frequent, and relatively
slow, it is important to keep the cpu humming by
working on several user programs concurrently.
Interleaving actions of different user programs
can lead to inconsistency: e.g., check is cleared
while account balance is being computed.
 DBMS ensures such problems don’t arise: users
can pretend they are using a single-user system.

Transaction: An Execution of a DB Program
Key concept is transaction, which is an atomic
sequence of database actions (reads/writes).
 Each transaction, executed completely, must
leave the DB in a consistent state if DB is
consistent when the transaction begins.




Users can specify some simple integrity constraints on
the data, and the DBMS will enforce these constraints.
Beyond this, the DBMS does not really understand the
semantics of the data. (e.g., it does not understand
how the interest on a bank account is computed).
Thus, ensuring that a transaction (run alone) preserves
consistency is ultimately the user’s responsibility!
Scheduling Concurrent Transactions

DBMS ensures that execution of {T1, ... , Tn} is
equivalent to some serial execution T1’ ... Tn’.



Before reading/writing an object, a transaction requests
a lock on the object, and waits till the DBMS gives it the
lock. All locks are released at the end of the transaction.
(Strict 2PL locking protocol.)
Idea: If an action of Ti (say, writing X) affects Tj (which
perhaps reads X), one of them, say Ti, will obtain the
lock on X first and Tj is forced to wait until Ti completes;
this effectively orders the transactions.
What if Tj already has a lock on Y and Ti later requests a
lock on Y? (Deadlock!) Ti or Tj is aborted and restarted!
Ensuring Atomicity
DBMS ensures atomicity (all-or-nothing property)
even if system crashes in the middle of a Xact.
 Idea: Keep a log (history) of all actions carried out
by the DBMS while executing a set of Xacts:



Before a change is made to the database, the
corresponding log entry is forced to a safe location.
(WAL protocol; OS support for this is often inadequate.)
After a crash, the effects of partially executed
transactions are undone using the log. (Thanks to WAL, if
log entry wasn’t saved before the crash, corresponding
change was not applied to database!)
The Log

The following actions are recorded in the log:

Ti writes an object: The old value and the new value.
• Log record must go to disk before the changed page!

Ti commits/aborts: A log record indicating this action.
Log records chained together by Xact id, so it’s easy to
undo a specific Xact (e.g., to resolve a deadlock).
 Log is often duplexed and archived on “stable” storage.
 All log related activities (and in fact, all CC related
activities such as lock/unlock, dealing with deadlocks
etc.) are handled transparently by the DBMS.

Databases make these folks happy ...
End users and DBMS vendors
 DB application programmers



E.g., smart webmasters
Database administrator (DBA)




Designs logical /physical schemas
Handles security and authorization
Data availability, crash recovery
Database tuning as needs evolve
Must understand how a DBMS works!
These layers
must consider
concurrency
control and
recovery
Structure of a DBMS



A typical DBMS has a
Query Optimization
layered architecture.
and Execution
The figure does not
Relational Operators
show the concurrency
Files and Access Methods
control and recovery
components.
Buffer Management
This is one of several
Disk Space Management
possible architectures;
each system has its own
variations.
DB
Database Management Systems 3ed, R. Ramakrishnan and J. Gehrke
38
Summary
DBMS used to maintain, query large datasets.
 Benefits include recovery from system crashes,
concurrent access, quick application
development, data integrity and security.
 Levels of abstraction give data independence.
 A DBMS typically has a layered architecture.
 DBAs hold responsible jobs
and are well-paid! 
 DBMS R&D is one of the broadest,
most exciting areas in CS.
