The
Transaction Processing Performance Council (TPC) today announced the
immediate availability of TPCx-HCI, the first application system-level
benchmark for measuring the performance of hyper-converged
infrastructure clusters. TPCx-HCI enables direct comparisons of price
and performance between hyper-converged infrastructure platforms, which
include hypervisors for virtualized computing, software-defined storage,
virtualized networking and commodity servers.
"When
compared to legacy solutions, hyper-converged infrastructure promises
both simplicity and flexibility," said Reza Taheri, chairman of the
TPCx-HCI committee and principal engineer at VMware. "At the same time,
creating a benchmark for hyper-converged systems and all of their
integrated components posed a significant challenge. I'm thrilled that
some of the best and brightest minds from leading competitor companies
in this market segment came together under the umbrella of the TPC, to
accomplish a common objective, enabling us to deliver this benchmark in
just 7 months."
Until
now, the only available benchmarks for the hyper-converged
infrastructure market have been the simple storage-only micro-benchmarks
that measure input/output operations per second (IOPS). TPCx-HCI
measures the performance of a complex application at the system level,
and was crafted in direct response to demand from the user community.
TPCx-HCI
stresses the virtualized hardware and software of converged storage,
networking, and compute platforms under a demanding database workload,
and has two primary unique characteristics:
- The
benchmark has an elastic workload that varies the load delivered to
each of the virtual machines (VMs) by as much as 16 times, while
maintaining a constant load at the cluster level. Sustaining optimal
throughput for this elastic workload on a multi-node HCI cluster would
typically benefit from frequent migrations of VMs to rebalance the load
across nodes. This property measures the efficiency of VM migration as
well as the uniformity of access to data from all the nodes.
- In
the Data Accessibility test of TPCx-HCI, a node is powered down
ungracefully, and the benchmark continues to run on the other nodes. The
test sponsor must disclose a throughput graph for this test,
demonstrating the impact on performance, as well as report the recovery
time to regain resilience.
Additionally, as an "Express" class benchmark, TPCx-HCI is available for download via the TPC's Web site: http://www.tpc.org/tpcx-hci/default.asp.
Organizations interested in contributing to the TPC's benchmark
development process are also encouraged to become members, and
additional information is available via the following URL: http://www.tpc.org/information/about/join.asp.
TPC
members that contributed to the development of TPCx-HCI include Cisco,
DataCore, Dell, HP Enterprise, Huawei, Microsoft, Nutanix, Oracle, Red
Hat and VMware.