Technologies Paper

Table 1 provides a sample four-year
operational cost comparison between a
1000-node Intel Xeon processor E5-2680
server cluster, and a 334-node cluster
based on Intel Xeon processor E7-4890.
Using the average of the Sort, TeraSort,
K-means, Apache Hive Join, and Page Rank
benchmarks on the 1.5 TB conguration as
a baseline, a 334-node cluster can do the
same amount of processing as the 1000-
node cluster. The initial acquisition costs for
both clusters vary based on the OEM and
conguration, but the operational costs
over a four-year period are considerably
less for the 334-node cluster.
2
By deploying higher performance servers
with the Intel Xeon processor E7 v2
family, enterprises can scale up their
Apache Hadoop clusters by using fewer,
more powerful servers. A scale-up
architecture reduces cluster complexity
and can reduce operational costs, which
can lead to lower TCO.
TABLE 1: FOUR-YEAR OPERATIONAL COSTS COMPARISON BETWEEN A 1000-NODE INTEL XEON PROCESSOR E5-2680 CLUSTER
AND A 334-NODE INTEL XEON PROCESSOR E7-4890 V2 CLUSTER.
ANNUAL COSTS PER SERVER
2
MAINTENANCE
POWER AND
COOLING YEARS
1 THROUGH 4 RACK SPACE
NETWORK
MAINTENANCE TOTAL COSTS
2-socket Intel® Xeon® processor
E5-2680 (2.7 GHz, 8 cores)
$500 $496 $155 $15 $1166
4-socket Intel® Xeon® processor
E7-4890 v2 (2.8 GHz, 15 cores)
$500 $1443 $310 $30 $2283
TOTAL 4-YEAR CLUSTER COST AT
EQUIVALENT PERFORMANCE
1000-node cluster with Intel
Xeon processor E5-2680
$2,000,000 $1,984,185 $620,000 $60,000 $4,664,185
334-node cluster with Intel Xeon
processor E7-4890 v2 servers
$668,000 $1,835,087 $414,160 $40,080 $2, 957,327
CLUSTER COST SAVINGS USING
SERVERS BASED ON THE INTEL
XEON PROCESSOR E7 V2 FAMILY
67
percent
8
percent
33
percent
37
percent
37
percent overall
cost savings
Scale Up with Intel Technologies
Many enterprises have built their Apache
Hadoop clusters on lower-performing
commodity hardware. Yet advances in
Intel CPU, networking, and storage
technologies can greatly enhance
the performance of big data analysis,
especially in large Apache Hadoop
clusters. These technologies include:
• The Intel Xeon processor E7 v2 family,
which provides up to 15 cores per
socket and up to 12 terabytes of RAM
in an eight-socket conguration for
faster access to data and greater
processing speed. New RAS features
can also increase the reliability of
Apache Hadoop clusters.
• Intel SSDs, which provide greater
reliability and performance that is
signicantly higher than traditional
mechanical hard disks.
• 10 gigabit Intel Ethernet Server
Adapters, which can increase the
networking throughput across Apache
Hadoop clusters while decreasing port
count and cabling complexity.
For more information on how
you can increase your enterprise’s
big data performance, go to
www.intel.com/hadoop.
6
Accelerate Big Data Analysis with Intel® Technologies