Fusion IO Blows Through 1 million IOPS to Nine Million IOPS in Transaction Logging App Use-Case

Everywhere there is this need for speed. One flash array vendor thought that benchmarking a single VM to achieve a million IOPS was particularly useful or interesting. Marginally.  However, I found that the real example of speed comes from Fusion IO and it has demonstrated it could deliver 9 million IOPS from a single 365 GB MLC Fusion ioDrive2 card. Watch the video the number is actually 9.6 million IOPS.  Unbelievable.


And a quick brief on the use-case :


Here is a nice detailed presentation/talk from youtube on it :

And a presentation on it :


The key aspect beyond the sheer hardware effort is the work done in building programming interfaces via their SDK.  By focusing not simply on sheer hardware but also on building useful programming interfaces for software developers, Fusion IO provides the tools to achieve very high IOPS.  Many flash companies have focused on hardware – but here Fusion IO shows that a focus on hardware and software and file systems is at the heart of their tremendous performance.


Recommended: Flash Accelerating Oracle RAC in 16RU & VMware Power Savings with Local Flash

Fusion-io, Dell, and Mellanox have built a very small footprint (16RU), extremely fast, and price-sensitive reference architecture for Oracle RAC 12c. It is very impressive on a number of counts.  It further highlights the move to using local flash storage over remote flash storage arrays.  This solution aims at reducing hardware sprawl and power consumption by achieving higher performance in a smaller footprint.  From the description – “This reference architecture features four Fusion ION Accelerators and four Oracle RAC database nodes all connected through redundant Mellanox SX6036 switches. Each ION Accelerator in this reference architecture uses a Dell PowerEdge R720 with three 56 Gbps Mellanox ConnectX-3 InfiniBand cards and four industry-leading Fusion ioDrive2 2.4TB flash storage devices for a total of 9.6TB of all flash storage. Each of these ION Accelerator units fits in 2 rack units (2RU) and delivers well over 645,000 8K database IOPS and up to 12 GB/s sustained throughput. Data redundancy is maintained across pairs of ION Accelerators using synchronous writes, thus providing system and data high availability.  The Oracle RAC nodes consist of Dell PowerEdge R620 servers. Each two-socket server consumes a mere 1U of rack space and yet is capable of pulling 1.4 million 8K IOPS from the ION Accelerator storage layer as measured by the Flexible IO Tester utility.  The Oracle RAC nodes are connected to the ION Accelerators through redundant Mellanox switches. Each Oracle RAC node sees the ION Accelerator storage as simple multipath block storage. The multipath devices are aggregated by Oracle ASM to create a large and powerful diskgroup.  Expanding the size and performance of the database is as easy as adding more ION Accelerator devices to the ASM disk group.”  The links to the article and the Reference Architecture follow.



Increasingly, software vendors  see the virtue of putting their flash storage locally on the server.  For example, VMware engineers wrote recently that by replacing array shared storage with PCIe flash card storage on the servers they could substantially reduce power consumption and maintain high levels of performance.




Fusion IO Flash Storage Accelerates In-Memory SQL Server 2014 Database up to 4x

Fusion IO has jumped on-board the Microsoft SQL Server 2014 in-memory features and has claimed up to 4x improvements in transaction per second and substantial reduction in data latencies.  A nice little Fusion IO video that introduces SQL Server 2014:

Fusion ioMemory integrates into the SQL Server 2014 database engine buffer pool to improve IO throughput. The lower price points of the PCIe flash card products will no doubt put pressure on flash arrays.


Fusion IO has written a nice white paper detailing how their cards accelerate SQL Server 2014 by leveraging the new In-Memory features. Using their ioDrive2 Duo cards they substantially accelerated the performance, latency of transactions.

fio101The performance brief shows an increase of 4.4 x more transaction with the ioDrive2 Duo 2.4TB versus an enterprise class disk array. It also shows that it delivers better latency – serving customers up to 73% faster and delivers 3.3x faster database startup times.


Here is a SQL Server 2014 In-Memory OLTP Overview :


Cisco Using Fusion-IO’s ioDrive2 Achieves TPC-H #1 Result

Following on the excellent results of Micron’s P320h we see that Fusion-IO is not sitting still.  A new TPC-H SAP Sybase IQ Benchmark offering up the best result by using Fusion-IO ioDrive2.  The Cisco UCS C420 M3 Rack Server used six Fusion ioDrive2 cards to achieve the result.



In flash PCIe cards, Fusion IO has developed an excellent record which they are legitimately promoting :

See also Six World Records and Nine of the Top Ten VMmark Scores Include Fusion-io

gotostorageGo to more posts on storage and flash storage at http://digitalcld.com/cld/category/storage.

Recommended Viewing : Why Micron is 5 Times Faster Than Fusion-IO

This is a particularly interesting for those working with NoSQL databases. In a meet-up hosted by Aerospike some of results from recent testing were provided.  In the beginning there is a discussion by Brian Bulkowski, CTO of Aerospike who talks about why they are interested in fast flash implementation.  I wrote about this recently in the post, Aerospike Benchmark :  Micron PCIe P320h and P420hm Flash Cards “Blow Away The Competition. They recently benchmarked Micron’s PCIe flash cards and were stunned by the performance – it was considerably faster than every other competing implementation.  Scott Shadley, Jr. from Micron presented how Micron is able deliver the performance that surprised Aerospike and also covered the 2.5″ PCIe P320h form-factor that I discussed in the posting, Micron’s P320h 2.5″ Flash PCIe Form-Factor Offers Hot-Swap, which I recommended anyone buying a PCIe flash card should look at.



gotostorageGo to more posts on storage and flash storage at http://digitalcld.com/cld/category/storage.


Recommended Reading : Reference Architecture for Oracle RAC with Fusion IO’s ION Data Accelerator Software and Dell r720 Servers

One basic question that is often asked is – should I use PCIe-based Flash Card or Flash/SSD array? Flash/SSD arrays are being used within enterprise and cloud settings but a considerable number of companies are also using PCIe-based flash cards inside their servers to run databases and applications.  The dominant PCIe vendor is clearly Fusion IO. Hands down.  You begin to see why when you see how much time and energy they have put into the software side of their product as well as hardware.   They have targeted some use-cases which make perfect sense for using PCIe flash cards within a cloud – for example Facebook is running MySQL on servers with PCIe flash cards.  Apple also is using PCIE cards in conjunction with database in iCloud.  Specifically around the MySQL ecosystem – Fusion IO is has been working with SkySQL to get atomic writes in MariaDB (a MySQL fork) and into the Percona Server.  This optimization on MariaDB has yielded over 50% improvement in write performance.   It is this type of value that differentiates vendors and is missing from some of the other PCIe vendors.  Another excellent use-case is around NoSQL databases like Aerospike. It is such an excellent use case that they have created a benchmark and they recommend SSDs or other flash devices.

One very interesting read is the Fusion IO whitepaper describing Reference Architecture for Oracle RAC with ION Data Accelerator and Dell r720 Servers. It shows a resilient architecture for running Oracle RAC.  The aim of this architecture is to reduce hardware costs and per-core Oracle licensing by 50% compared to an Exadata X3-2 Half Rack.  At the same time, there is a reduction of power consumption and delivering high IOPS for this architecture. When compared to an Oracle Exadata X3-2 the Fusion IO/Dell architecture consumes only a quarter rack (compared to half), 32 instead of 64 cores, requires only two database servers (compared to 4), requires only two storage servers compared to seven and provides 800,000 Read Only IOPS (compared to Exadata’s 750,000 IOPS) and provides a 50% cost reduction in licenses.  Pretty worthwhile reading is Fusion IO’s whitepaper on how they do this :