Data in your SOA: From SQL to NoSQL and Beyond
-
Upload
wso2 -
Category
Technology
-
view
1.588 -
download
3
Transcript of Data in your SOA: From SQL to NoSQL and Beyond
1
In WSO2Con 2010….
2
3
topic: Unleash the power of enterprise data
4
power of data
5
data centricity
6
7
why
8
how
9
10
this year …
11
topic : Data in your SOA -‐ SQL to NoSQL
12
storage options
13
selection
14
“Data in your SOA : SQL to NoSQL”
15
data
16
located in
17
18
has
19
attributes
20
<attributes>
21
#1. lifecycle
22
#2. owner (provenance)
23
#3. access pattern
24
#4. structure
25
#5. no-‐structure
26
#6. semi-‐structure
27
</attributes>
28
some examples..
29
CNN iReport
30
CNN iReport
Multi media
linked
tagged
News items
searched
access patterns
31
32
user accounts
relationships
images
discussions
33
Point of Sales App
34
Point of Sales App
item list
mode of payment
35
Amazon.com
36
Amazon.com
Products
Suppliers
Orders
History relationships
37
middleware app
38
middleware app
pass thru msg
meta data
39
con]igurations
how data are stored?
40
or, how would you be doing it?
41
<expectations>
42
reliable storage
43
retrieval
44
update
45
non-‐functional expectations
46
</expectations>
47
options
48
49
<options>
50
relational databases
moderate scaling read/write/search
transactional de facto standard
51
spreadsheets
familiarity
built-‐in functionality
single user
52
online spreadsheets
multi user
53
API
Google Fusion Tables
54
message queues
synchronize processing speed
transactional
temporarily storage
reliable
55
registries
specialized APIs
manageability
56
]ile system
large ]iles
transactional support
57
cache(local/distributed)
fast no persistence
memory
key no search
58
NoSQL column family
59
why?
60
Internet
61
applications
62
scaling
63
users
64
Big Data
65
nosql-‐database.org (122+)
66
document databases
67
key/value storages
68
object databases
69
XML databases
70
many more...
71
</options>
72
different data models
73
how to select?
74
things to consider…
75
questions to ask…
76
<questions>
77
data type
78
<types>
79
structured
80
[[example]]
81
un-‐structured
82
[[example]]
83
semi-‐structured
84
<types>
85
volume & growth
86
scalability
87
#1: vertical
88
#2: horizontal
89
<horizontal>
90
eg: MySQL mirroring/clustering
91
limitations
92
CAP theorem
93
consistency
Partition tolerance
availability
</horizontal>
94
data access patterns
95
<patterns>
96
archival
97
Analytics (OLAP)
98
OLTP
99
</patterns>
100
<retrieval>
101
#1: sequential
102
#2: key based
103
supports : un-‐structured, structured
104
#3: query based
105
supports : structured
106
</retrieval>
107
life span
108
transaction support
109
manageability
110
</questions>
111
<some guidelines>
112
transaction intensive
113
small scale,
114
JOINs
115
relational databases.
116
un-‐structured data,
117
no search
118
key/value storages
119
or column family.
120
potential for SCALING
121
writes && reads
122
relational databases
123
not the best choice.
124
hybrid
125
approaches.
126
build
127
expertise.
128
your data
129
data model
130
storage engine.
131
Don’t
132
COP/ROP.
133
</some guidelines>
134
WSO2 platform
135
on premise
136
or Java PaaS
137
applications
138
need
139
data storage
140
can make use of
141
WSO2 Message Broker (MB)
142
WSO2 Relational Storage Service (RSS)
143
WSO2 Cassandra Storage Service (CSS)
144
WSO2 Governance Registry
145
WSO2 CSG
146
more on line up
147
interested?
148
tutorial session
149
problems !!
150
heterogeneous storages
151
in summary…
152
no silver bullet
153
analyze, choose right one
154
thank you.
155