HP Integrity rx1620 Server 1.30 GHz Windows Developers Bundle AB430A#007 プリント

Data protection 
Event data must be protected from being changed or destroyed by any source, including applications, 

users, and component failures. Destruction or modification of event data can result in non-compliance 
with government regulation, increased security risk, or failures in system management. Although 

relational data has similar data protection requirements, in many cases these requirements are not as 

absolute, and protection failures do not have the same pervasive impact. 

Table1: Contrasting relational data with event data 

Area of concern 

Event data 

Relational data 

Transactions 

Limited transactional requirements Fundamental 

requirement 

Stored data is seldom updated so all stored data is 
available to all users (subject to authorization 
filtering). 

Must present an isolated virtual database 
view to prevent visibility of non-committed 
data. 

Isolation concurrency 

Must be general and flexible to accommodate 
future event types. Semantics of data are often 
determined at search time. 

Generally static and must be determined 
before data can be stored and accessible 
by applications. 

Schema 

Search criteria are precise and databases 
are optimized to support a known set of 
queries. Search requirements are static or 
evolve slowly. 

Search criteria can be precise or pattern based and 
search requirements evolve rapidly. 

Search 

Time attribute is part of every event and is usually a 
key criterion for search. 

Time is one of many possible attributes 
and may not be present. 

Time attribute 

It has no consistent life cycle 
requirements. Most data has an indefinite 
life span. 

All event data is eventually removed or archived 
based on aging and retention rules. 

Life cycle 

Data protection to prevent destruction and 
modification is absolute and must be supported 
regardless of use access or component failure. Data 
protection failures may cause compliance failures. 

Data protection is customized per 
application and is based on database 
security authorizations and application 
business logic. 

Data protection 

Lack of availability is likely to impact the entire 
system and may have legal ramifications. 

Lack of availability may only impact a 
segment of the system. 

High availability 

Load rate must keep pace with the event data 
creation rate. System components generating event 
data do not wait for event data loading. 

Load rate is based on user response time. 
Increased response time reduces the load 
rate requirements. 

Load rate 

 

Storage barriers of RDBMS-managed event data 

The mismatches between event data management requirements and RDBMS core technology strengths 

can be roughly depicted as either: 

•  RDBMS over-head is not required to manage event data or  
•  Lack of technology to support the unique requirements of event data management. 

Industry experience and empirical evidence show that these differences are manifested in real failures 

to meet compliance, security and system management requirements. These failures can result in failing 

to meet audits, increased security risks, and increased IT infrastructure costs. 
There are multi-faceted problems in using an RDBMS to manage event data. Companies using RDBMS 

technology in this manner are likely to experience most, if not all, of these problems. 

7