Databases Reference
In-Depth Information
There are several variants of in-memory architectures, and we will discuss briefly the common
goals and benefits of in-memory computing as applied to Big Data and data warehouse solutions.
In-memory technologies enable storage of vast amounts of data in an online storage and reduce the
roundtrips between disk and memory, circumventing latencies to near zero. The ability to load and
unload data at extremely high speeds in these architectures provides architects with options to create
platforms for high-speed data movement and management.
Benefits of in-memory architectures
●
Agility of analytics and number-crunching operations
●
Lookup and MDM data management without disk operations
●
Stream analytics implementation
●
Inspect data on the wire before it reaches disk
●
Persist standard reports and precalculated data in memory
●
Near-real-time refresh
●
Drill down, drill across, and roll up with extreme flexibility
Evolving areas of in-memory computing:
●
Data security
●
Scalability
●
Compression
While there are several in-memory technologies and solutions available today, the user adoption
is not as viral, and the reason for this is too much dependency on legacy solutions, fear of moving to
in-memory computing, uncertainty of sustained performance, and user reluctance to move from an
incumbent platform.
the most popular infrastructure solutions include Fusion IO; Violin Memory; business intelligence
platforms like Qlikview, SAP, and HANA; and data warehouse platforms like MongoDB, Oracle,
IBM, and Microsoft with Fusion IO or Violin Memory.
SUMMARY
This chapter provided some of the groundbreaking options in terms of infrastructure, where the data
architecture and workload management capabilities combined together enable a robust infrastructure
for the next-generation data warehouse. What exactly is the architecture for the next-generation data
warehouse? How can we physicalize the infrastructure while retaining the foundation goals of the
data warehouse? These questions are the focus of discussion in the next chapter.
Further reading
