The Data Distribution Service (DDS) is a standard for ubiquitous, interoperable, secure, platform independent, and real-time data sharing across network connected devices. DDS is today used in a large class of applications, such as, Power Generation, Large Scale SCADA, Air Traffic Control and Management, Smart Cities, Smart Grids, Vehicles, Medical Devices, Simulation, Aerospace, Defense and Financial Trading.
Differently from traditional message-centric technologies, DDS is data-centric – the accent is on seamless (user-defined) data sharing as opposed to message delivery. Therefore, when embracing DDS and data-centricity, data modeling becomes a key step in the design of a distributed system.
This webcast will (1) explain the role and scope of data modeling in DDS, (2) introduce the techniques at the foundation of effective and extensible Data Models, and (3) summarize the most common DDS Data Modeling Idioms.
Six Myths about Ontologies: The Basics of Formal Ontology
The Art and Science of DDS Data Modelling
1. The Art and Science of DDS
Data Modelling
Angelo Corsaro, PhD
Chief Technology Officer
OMG DDS SIG Co-Chair
angelo.corsaro@prismtech.com
PrismTech
2. A Recurring Question
• People new to DDS recurrently ask a question: what are the techniques and
• My answer is usually: Start with the powerful tools and techniques provided
by relational data modelling and then add some DDS-specific spice
• I’ve come to the conclusion that many people are not very familiar with
relational data modelling, or perhaps it is way too long that they have
studied/reviewed these concepts
• This webcast, will provide a relatively well introduction to the relational
data model
PrismTech
Copyright PrismTech, 2014
patterns that we can use to design DDS-based Systems?
4. • Introduced by Edward Codd in 1970 as a way of representing data models
for Data Bases
• Simple and Elegant: A database becomes a collections of one or more
relations where each relation is a table with rows and columns
PrismTech
Copyright PrismTech, 2014
Relational Model
5. Relation
it consists of two dimensional table
• The columns of a relation are called attributes
• The name of the relation along with the set of attributes defines the relation
schema
• The rows of the relation, other than the header containing the attribute
names, are called tuples
PrismTech
Copyright PrismTech, 2014
• The relation is the construct used representing data in the relational model,
6. Relation’s Schema
-
relation’s name
-
the name of each field/attribute, e.g. column
-
the domain of each field, e.g. the type of the field
!
• Example:
-‐
PrismTech
Student(sid:
string,
name:
string,
age:
integer,
gpa:
real)
Copyright PrismTech, 2014
• The relation schema specifies:
7. Tuples
• An instance of a relation is a set of tuples (records) in which each tuple has the same
number of fields as in the relation schema.
rows have the same number of fields (columns)
!
sid
!
!
!
name
age
gpa
1234
Peter Parker
21
4.0
2345
3456
Tony Stark
Bruce Wayne
15
23
4.0
3.5
• Notice that rows are all different. This is a requirement of the relational model, as a
relation instance is a collection of unique tuples (or rows)
PrismTech
Copyright PrismTech, 2014
• A relation’s instance can be visualised as table where each tuple is a row and all
8. • The cardinality of a relation R is defined as the number of tuples belonging
to the relation
• The degree, or arity, of a relation R is defined as the number of its fields
PrismTech
Copyright PrismTech, 2014
Cardinality and Degree
9. Keys
• A superkey is a set of attributes that includes the primary key
• Example:
-
The sid field is the key for the Students relations
sid
1234
2345
3456
PrismTech
name
Peter Parker
Tony Stark
Bruce Wayne
age
21
15
23
gpa
4.0
4.0
3.5
Copyright PrismTech, 2014
• The key of a relation is a set of fields that uniquely identifies a tuple
10. • A foreign key allows to introduce a link between two relations
• For instance, the sid in the Courses relation is a foreign key allow to refer as
well as introduce an integrity constraint to the students relations
Courses
cid
sid grade
Physics303 1234 A+
Robotics323 2345 A+
Calculus343 2345 A
PrismTech
Students
sid
1234
2345
3456
name
Peter Parker
Tony Stark
Bruce Wayne
age
21
15
23
gpa
4.0
4.0
3.5
Copyright PrismTech, 2014
Foreign Keys
12. • DDS provides a Global Data Space
abstraction that allow applications
to autonomously, anonymously,
securely and efficiently share data
• DDS’ Global Data Space is fully
distributed, highly efficient and
scalable
PrismTech
Copyright PrismTech, 2014
Data Distribution Service (DDS)
13. • DataWriters and DataReaders are
automatically and dynamically
matched by the DDS Discovery
• A rich set of QoS allows to control
existential, temporal, and spatial
properties of data
PrismTech
Copyright PrismTech, 2014
Data Distribution Service (DDS)
15. Topic
Copyright PrismTech, 2014
• A Topic defines a domain-wide information’s class
• A Topic is defined by means of a (name, type,
qos) tuple, where
-
type: is the programming language type associated
with the topic. Types are extensible and evolvable
Name
-
qos: is a collection of policies that express the nonfunctional properties of this topic, e.g. reliability,
persistence, etc.
Topic
e
PrismTech
Qo
S
name: identifies the topic within the domain
Typ
-
16. Topic and Instances
• As explained in the previous slide a topic defines a class/type of information
• Topic Instances are identified by means of the topic key
• A Topic Key is identified by a tuple of attributes -- like in databases
• Remarks:
-
PrismTech
A Singleton topic has a single domain-wide instance
A “regular” Topic can have as many instances as the number of different key
values, e.g., if the key is an 8-bit character then the topic can have 256 different
instances
Copyright PrismTech, 2014
• Topics can be defined as Singleton or can have multiple Instances
17. • IDL is the most commonly used syntax
Topic
• Example:
struct
Student
{
long
sid;
string
name;
int
age;
float
gpa;
};
#pragma
keylist
Student
sid
PrismTech
Qo
S
Name
e
Typ
• A Topic type can be defined in different syntaxes
Copyright PrismTech, 2014
Topic Example
18. Topics as Relations
struct
Student
{
long
sid;
string
name;
int
age;
float
gpa;
};
#pragma
keylist
Student
sid
Student(sid, name, age, gpa)
name
age
gpa
1234
Peter Parker
21
4.0
2345
3456
PrismTech
sid
Tony Stark
Bruce Wayne
15
23
4.0
3.5
Copyright PrismTech, 2014
• A Topic cans be seen as defining a relation
19. • Topics Types => Relation Schema
• Topic Instance => Key
• Topic Sample => Tuple
PrismTech
Copyright PrismTech, 2014
Mapping DDS to the Relational Model
20. • Start identifying corse relations and properties of data
• Start decomposing based on properties
• Apply a normal form
-
PrismTech
Functional Dependencies => Boyce-Codd Normal Form
Multivalued Dependencies => Fourth Normal Form
Copyright PrismTech, 2014
Relational Design
22. UML Data Modelling
• A subset of UML can be used to model Data Models
• The resulting model can be easily translated into a relational model and the used
• The allowed subset of UML are:
-
Classes (with only attributes)
-
Associations
-
Association Classes
-
Subclasses
-
Composition and Aggregation
• UML Data Models can be automatically translated into relational model as far as
each “regular” class defines a primary key
PrismTech
Copyright PrismTech, 2014
in a DBMS or DDS
23. Class
• A UML class is mapped to a relation that has the same name of the class,
Student
sid: int
name: string
age: int
gpa: float
PrismTech
Copyright PrismTech, 2014
shares its key and attributes
Student(sid, name, age, gpa)
24. • By default association can be mapped as follows, yet, depending on the
multiplicity of the association different mappings may be possible/desirable
!
!
C1
K1: PK
O1
A
C2
K2: PK
O2
C1(K1, O1)
C2(K2, O2)
A(K1,K2)
!
• The key definition in the association depends on the multiplicity
PrismTech
Copyright PrismTech, 2014
Association
25. 1-to-many Association
M1 Use a relation to capture the association
M2 Embed the association on the many side of the association
C1
K1: PK
O1
PrismTech
0..1
A
C2
* K2: PK
O2
M1 C1(K1, O1), C2(K2, O2), A(K1, K2)
M2 C1(K1, O1), C2(K2, O2, K1)
Copyright PrismTech, 2014
There are two ways of mapping a 1-to-many association to the relational
model
28. C1
K1: PK
O1
A
C2
K2: PK
O2
C1(K1, O1)
C2(K2, O2)
Association
A
PrismTech
A(K1,K2, a1, a2)
Copyright PrismTech, 2014
Association Classes
29. tsdotd14
Self Association
• Self association are modelled as traditional relations, which the only
Student
sid: int
name: string
age: int
gpa: float
*
PrismTech
Student(sid, name, age, gpa)
Sibling(sidParent, sidSibling)
*
Slbling
Copyright PrismTech, 2014
difference that attributes mau be conserved
30. Subclasses
Three ways of mapping subclassing to the relational model
T2 Subclass relations contain all attributes
T3 One relation containing all superclass and subclass attributes
A
K: PK
X
T1 A(K, X), B(K, Y), C(K, Z)
T2 A(K, X), B(K, X, Y), C(K, X, Z)
B
Y
C
Z
T3 A(K, X, Y, Z)
The best translation may depend on the the context, e.g. T3 good for heavily
overlapping subclasses, T2 good for disjoint and complete subclasses
PrismTech
Copyright PrismTech, 2014
T1 Subclass relations contain the superclass key and the specialised attributes
31. Composition and Aggregation
• The precondition to easily map composition to the relational model is for
Whole
K: PK
W
Part
P
Whole(K, W)
Part(P, K)
• When mapping aggregation (unfilled diamond), the key K on the Part
should have a domain that allows for null values
PrismTech
Copyright PrismTech, 2014
the part not to have a key
32. • A subset of UML can be used to model relational data models
• The mapping rules can be used to help translating existing Object Oriented
data models into their relational counter-part
PrismTech
Copyright PrismTech, 2014
Summing Up
34. Why Relation Refinement?
• The UML/ER Data Models provide usually a good starting point toward the
• The relations implied by the UML/ER Data Model often need to be
normalised and re-organised to address performances and workload criteri
• The goal of relation refinements is to remove redundancy and/or
decompose a relation with smaller relations
• Normal forms provide a way of measuring the amount of redundancy that
may be in our data model
PrismTech
Copyright PrismTech, 2014
data model that we’ll actually use in the system
35. Redundancy
• Redundant Storage: Information may be stored multiple times leading to
• Update Anomalies: If one copy of the redundant information is update this
may create inconsistencies in other copies — unless all copies are updated
at the same time
• Insertion Anomalies: It may not be possible to store some information,
unless some other information is stored as well
• Deletion Anomalies: It may not be possible to delete some information
without loosing som other information as well
PrismTech
Copyright PrismTech, 2014
space, and perhaps time, inefficiencies
36. Decomposition
• Unconsidered decomposition can lead more problems than benefits, thus
-
You really need to decompose the relation
-
You fully understand the implications of the decomposition (lossless join,
dependency preservation)
• Normal Forms provide good guidelines for relations decompositions as they
guarantees that certain class of problems cannot be introduced
• Notice that decomposition can have a performance impact as it may
lead to an increase in joins
PrismTech
Copyright PrismTech, 2014
when decomposing you always want to ensure that:
37. Functional Dependencies
• A Functional Dependency (FD) is a kind of Integrity Constraint (IC) that
• Given a relation R along with two nonempty sets of attributes X and Y in R,
we say that R satisfies the FD X ⟶ Y if the following holds for every pair of
tuples t1 and t2 in R:
!
if t1.X = t2.X then t1.Y = t2.Y
• In other terms, the FD says that if two tuple agree on the set of attributes on
X they also agree on the set of attributes in Y
• Notice that a primary key constraint is a special kind of FD
PrismTech
Copyright PrismTech, 2014
generalises the concept of a key
38. Example
percentile of the student GPA, e.g. which percentage of students has a GPA that is
smaller of equal
!
sid
!
!
name
age
gpa
percentile
1234
Peter Parker
21
4.0
100
2345
Tony Stark
3456 Bruce Wayne
15
23
4.0
3.5
100
75
!
• Clearly we have that the percentile attribute functionally depends on gpa, or
equivalently gpa ⟶ percentile
PrismTech
Copyright PrismTech, 2014
• Let’s assume our Student relation now includes a new attribute that measure the
40. Normal Forms
• Different Normal Forms (NF) exist that provide guidance on how to decompose
• If a relation is in a given normal form then we are guarantees that some
anomalies cannot arise, e.g. update anomaly, etc.
• The normal forms based on functional dependencies are the first normal form
(1FN), second normal form (2FN), third normal form (3NF) and the Boyce-Codd
normal form (BCNF)
• Every relation in BCNF is also in 3NF, every relation in 3FN is also in 2FN and finally
every relation in 2NF is also in 1NF
• The 2NF and 3NF have only historical interest, while the BCNF has important
practical applicability
PrismTech
Copyright PrismTech, 2014
relations
41. • A relation is in 1NF if every field contains only atomic values, that is not lists,
or sets
PrismTech
Copyright PrismTech, 2014
1NF
42. Boyce-Codd Normal Form (BCNF)
Let R be a relation, X a subset of attributes of R and a an attribute of R. R is in Boyce-Codd
Normal Form (BCNF) if for every FD: X ⟶ {a} that holds over R, one of the following is true:
Copyright PrismTech, 2014
• a ∊ X, that is it is a trivial FD, or
• X is a superkey
!
Intuitively, in a BCNF relation the only nontrivial dependencies are those in which a key
determines some attributes. Each attribute must describe the key, the whole key, and
nothing but the key
key
attr 1
attr 2
Functional Dependencies in BCNF
PrismTech
attr k
43. BCNF Decomposition Algorithm
Input: relation R and FDs for R
Compute Keys for R
Repeat until all relations are in BCNF
Choose a relation Ri with A ⟶ B that violates BCNS
Decompose Ri into R1(A, B) and R2(A, rest)
Compute FDs for R1 and R2
Compute Keys for R1 and R2
PrismTech
Copyright PrismTech, 2014
Output: decomposition of R into BCNF relations with lossless join
44. 3NF
• a ∊ X, that is it is a trivial FD, or
• X is a superkey, or
• a is part of some key for R
The definition of 3NF is similar to that of BCNF, with the difference that a may
be part of a key for R
PrismTech
Copyright PrismTech, 2014
Let R be a relation schema, X a subset of attributes of R and a an attribute of
R. R is in Third Normal Form if for every FD: X ⟶ {a} that holds over R, one of
the following is true:
45. Multivalued Dependencies
• For a relation R we say that A ↠ B (A multi-determines B), where A and B
!
!
!
∀ t,u ∈ R: t.A = u.A then ∃ v ∈ R:
v.A = t.A and
v.B = t.B and
v.rest = u.rest
!
• Multivalued dependencies are sometimes called tuple-generating
dependencies
PrismTech
Copyright PrismTech, 2014
are sets of fields in R, if:
46. • A relation R with multivalued dependencies (MVD) is in 4NF if for each nontrivial A ↠ B, A is a key
• The 4NF implies the BCNF
PrismTech
Copyright PrismTech, 2014
Fourth Normal Form (4NF)
47. 4NF Decomposition Algorithm
Output: decomposition of R into 4NF relations with lossless join
Compute Keys for R
Repeat until all relations are in 4NF
Choose a relation Ri with a nontrivial A ↠ B that violates 4NF
Decompose Ri into R1(A, B) and R2(A, rest)
Compute FDs and MVDs for R1 and R2
Compute Keys for R1 and R2
PrismTech
Copyright PrismTech, 2014
Input: relation R and FDs and MVDs for R
48. • Dependency enforcement may require joins
• Query workload — due to excessive joins
• Over-decomposition
PrismTech
Copyright PrismTech, 2014
Shortcomings of BCNF and 4NF
50. Selection and Projection
• Relational algebra provides operators to select rows (σ) an to project
columns from a relation (π)
Copyright PrismTech, 2014
• These operation allow to operate on a single relation
Examples:
σage<20 (Student)
Student
sid
name
1234 Peter Parker
age gpa
21
4.0
2345 Tony Stark
15
3456 Bruce Wayne 23
4.0
3.5
PrismTech
sid
name
2345 Tony Stark
age gpa
15 4.0
πname,gpa(Student)
name
gpa
Peter Parker 4.0
Tony Stark 4.0
Bruce Wayne 3.5
51. • Join is one of the most useful operator in relational algebra and is most
commonly used to combine/reassemble information from two or more
relations
• Join is conceptually a cross product followed by a selection and projection
PrismTech
Copyright PrismTech, 2014
Joins
52. • Condition joins are the most general form of joins. This operation takes a
condition and two relations and is defined as follows:
R ⋈c C = σc(RxS)
PrismTech
Copyright PrismTech, 2014
Condition Joins
53. • Equijoin is a special case of the Condition Join, where the condition
predicates on attribute equality
PrismTech
Copyright PrismTech, 2014
Equijoin
54. • A Natural Join is a special Equijoin that operates on all the attributes having
the same name in R and S
PrismTech
Copyright PrismTech, 2014
Natural Join
56. Relational Design in DDS
• Start decomposing based on properties (can use UML for this)
• Apply a normal form
-
Functional Dependencies => Boyce-Codd Normal Form
-
Multivalued Dependencies => Fourth Normal Form
• Define QoS for the resulting relations and further decompose if you incur in
some QoS Mix (more later)
PrismTech
Copyright PrismTech, 2014
• Start identifying corse relations and properties of data
57. • DDS Supports:
-
Selection for a given Topic DDS queries and filters
-
Conditional Joins across multiple Topics via the Multi-Topics
• DDS uses a subset of SQL-92 to express selections, projections and joins
PrismTech
Copyright PrismTech, 2014
Relational Algebra
58. • In some instances you may find that a topic (relation) R has two disjoint sets
of attribute X and Y that have conflicting temporal, reliability or durability
requirements
• In this case this relation has to be further decomposed
PrismTech
Copyright PrismTech, 2014
DDS Specific Decomposition
59. Frequency Mix
• Suppose you have a relation R(K, X,Y) were the set of attributes X changes
• In this case you should decompose the relation R into:
!
R1(K, X), R2(K, Y)
!
• This will reduce the resource usage in your system, e.g. bandwidth as well
as CPU but may introduce consistency issues. If consistency is essential then
coherent updates should be used to atomically update R1 and R2
PrismTech
Copyright PrismTech, 2014
far more frequently than the set of attributes Y (e.g. position, vs. velocity)
60. Reliability Mix
some soft-state.
• In this case you should decompose the relation R into:
!
R1(K, X), R2(K, Y)
!
• This decomposition allows to only use reliable distribution for R1 and besteffort for R2 thus reducing resource usage in the system
PrismTech
Copyright PrismTech, 2014
• Suppose you have a relation R(K, X,Y) were the set of attributes Y represent
61. Durability Mix
a different durability than the set of attributes Y, e.g. X need sto be
persistent while Y volatile
• In this case you should decompose the relation R into:
!
R1(K, X), R2(K, Y)
!
• This will reduce the resource usage in your system and reduce the pressure
on the Durability Service
PrismTech
Copyright PrismTech, 2014
• Suppose you have a relation R(K, X,Y) were the set of attributes X requires
63. Concluding Remarks
• The relational model provides the right set of tools for designing DDS-based
• DDS Topics are relations and DDS supports a subset of relational algebra to
manipulate these relations (topics)
• The design process is as follows:
-
Ensure your model is in BCNF or 4NF — make sure your understand why some
violations are necessary/desirable for your system
-
Add QoS to your relations
-
PrismTech
Start modelling your system using the UML Data Modelling subset
Evaluate if further decomposition is required due to QoS mixes — if your data
model is properly normalised
Copyright PrismTech, 2014
systems
65. • A First Course in Database Systems (3rd edition), Ullman and Widom
• Database Management Systems (3rd edition), Ramakrishnan and Gehrke
PrismTech
Copyright PrismTech, 2014
Books
66. • Jennifer Widom, Stanfords, Introduction to Databases
-
PrismTech
A very very good course on Databases in general and specifically on relational
data modelling
Copyright PrismTech, 2014
coursera.org
69. • Relational Data Models are commonly expressed using, some variation of,
Entity-Relationship (ER) Data Models
• The ER Data Model is built around the concepts of entities, attributes and
relationships (not to be confused with relations!)
PrismTech
Copyright PrismTech, 2014
Entity Relationship(ER) Data Model
70. Entities, Attributes and Entity Sets
• An entity is an object in the real world that is distinguishable from other
-
e.g. the iPhone, the Samsumg Galaxy Note, etc.
• An entity is described through a set of attributes
• An entity set identifies a collections of similar entities
-
e.g., Mobile Phones
• Each attribute associated with an entity set must identify its domain
• An entity has a primary key and potentially several candidate keys
PrismTech
Copyright PrismTech, 2014
objects
72. • A relationship is an association between two or more entities
-
e.g., a student is enrolled in a course
• A relationship can have descriptive attribute to record information about a
relationship
PrismTech
Copyright PrismTech, 2014
Relationships
73. Mapping
• The attributes of the resulting relation are:
-
the primary key of each participating entity as foreign keys
-
descriptive attributes as fields of the relation
• The primary key of the resulting relations depends on arity of the
relationship
PrismTech
Copyright PrismTech, 2014
• A relationship Set is mapped to a relation
74. Entity Hierarchies
• In some cases it is natural to introduce (type)
name
Copyright PrismTech, 2014
ssn
Employees
hierarchies among entities
• These hierarchies are represented through
the ISA relationship
hoursWorked
ISA
contractId
hourlyWages
HourlyEmpls
PrismTech
ContractEmpls
75. ISA relationships can be mapped into two ways
• Map each entity to a distinct relation
• Create only relations for the concrete types
Notice that while the first approach is always applicable, the second is not
PrismTech
Copyright PrismTech, 2014
Mapping