2 cissp ® Official Study Guide Eighth Edition

Large-Scale Parallel Data Systems
Parallel data systems
or 
parallel computing
is a computation system designed to perform 
numerous calculations simultaneously. But parallel data systems often go far beyond basic 
multiprocessing capabilities. They often include the concept of dividing up a large task into 
smaller elements, and then distributing each subelement to a different processing subsystem 
for parallel computation. This implementation is based on the idea that some problems 
can be solved efficiently if broken into smaller tasks that can be worked on concurrently. 
Parallel data processing can be accomplished by using distinct CPUs or multicore CPUs, 
using virtual systems, or any combination of these. Large-scale parallel data systems must 
also be concerned with performance, power consumption, and reliability/stability issues.
Within the arena of multiprocessing or parallel processing there are several divisions. 
The first division is between 
asymmetric multiprocessing (AMP)
and 
symmetric multipro-
cessing (SMP)
. In AMP, the processors are often operating independently of each other. 
Usually each processor has its own OS and/or task instruction set. Under AMP, proces-
sors can be configured to execute only specific code or operate on specific tasks (or specific 
code or tasks is allowed to run only on specific processors; this might be called 
affinity
in 
some circumstances). In SMP, the processors each share a common OS and memory. The 
collection of processors also works collectively on a single task, code, or project. A varia-
tion of AMP is massive parallel processing (MPP), where numerous SMP systems are linked 
together in order to work on a single primary task across multiple processes in multiple 
linked systems. An MPP traditionally involved multiple chassis, but modern MPPs are com-
monly implemented onto the same chip.
The arena of large-scale parallel data systems is still evolving. It is likely that many man-
agement issues are yet to be discovered and solutions to known issues are still being sought. 
Large-scale parallel data management is likely a key tool in managing big data and will 
often involve cloud computing, grid computing, or peer-to-peer computing solutions. These 
three concepts are covered in the following sections.
Distributed Systems and 
Endpoint Security
As computing has evolved from a 
host/terminal model
(where users could be physically 
distributed but all functions, activity, data, and resources reside on a single centralized 
system) to a 
client-server model
(where users operate independent, fully functional desktop 
computers but also access services and resources on networked servers), security controls 
and concepts have had to evolve to follow suit. This means that clients have computing and 
storage capabilities and, typically, that multiple servers do likewise. The concept of a client-
server model network is also known as a distributed system or a distributed architecture. 
Thus, security must be addressed everywhere instead of at a single centralized host. From a 
security standpoint, this means that because processing and storage are distributed on mul-
tiple clients and servers, all those computers must be properly secured and protected. It also 

---

Distributed Systems and Endpoint Security 
351
means that the network links between clients and servers (and in some cases, these links 
may not be purely local) must also be secured and protected. When evaluating security 
architecture, be sure to include an assessment of the needs and risks related to distributed 
architectures.