Posts Tagged ‘latency’
Posted April 17th, 2013
 |
Delivering on Our Vision for Enterprise Storage
By Dani Golan
|
Today is an important day, not just for Kaminario but, we believe, for the entire enterprise storage industry. Today is the day we announce the Kaminario fourth-generation K2 all-flash storage array. What is so important about our fourth-generation product? Here are a few key points:
Breakthrough TCO: With K2 v4 we have been able to increase the density by more than 500% while cutting the price in half. Why is this so significant? Lowering the cost of enterprise-ready solid-state flash storage has been one of the long-time goals of the industry. If you believe, as we do, that solid-state is the future of Tier One storage, then it must be cost-efficient for all applications, not just those requiring the highest levels of performance. And now it is. What’s more, Kaminario’s TCO breakthrough extends the viability of solid-state to a new class of mid-size customers who can now afford unprecedented levels of performance, resiliency and ease of use from their storage infrastructures.
Killer Performance: Even before today’s announcement Kaminario delivered by far the best performance in the industry, with world record SPC-1 results for sustained performance and price/performance. With K2 v4 we now have 400% more read/write bandwidth than our previous versions, with consistently low latency (120-microsecond writes). Topping it off, we still offer industry-leading IOPS performance. K2 v4 delivers significant across-the-board performance advantages over any other solution for any combination of workloads, be they OLTP, OLAP or virtualization. It’s a level of performance consistency that is unparalleled in the industry.
Bulletproof Resiliency: Our proven SPEAR Scale-Out Architecture maintains data integrity through any type of failure. For most workloads the performance levels of our solutions will be minimally impacted—usually less than 10%. We even guarantee that degradation will max out at 25%. What’s more, our snapshots are the most efficient in the storage industry, enabling instant restore and recovery from any snapshot with no impact on the performance of the production environment. Snaps can be taken in just milliseconds, 20 times faster than any legacy SAN system on the market today. We make snapshots sexy.
These are major advances in solid-state that have been more than five years in the making. They move storage technology to a whole new level. Don’t just take our word for it: Take a look at the new architectural white paper or better yet, the independent report.
On a personal note, it is an important day for me, as well as for the entire team here at Kaminario. From the beginning we envisioned a new era in enterprise storage driven by solid-state flash and a true scale-out architecture. As of today, that vision is a reality. What’s next? We can’t wait to tell you…
Tags:application performance, Database Performance, Flash, Flash SSD, IOPS, K2, Kaminario, latency, OLAP, OLTP, reads, Solid State Storage, SSD, Storage Performance, virtualization, writes
Posted in All Flash Storage Array | No Comments »
Posted March 25th, 2013
|
Good Riddance Indeed to the Spinning Disk Era
INFOWORLD AUTHOR LAMENTS USE OF ANCIENT STORAGE TECH
By Kaminario
|
Earlier this year, InfoWorld’s Paul Venezia penned Good-bye – and good riddance – to spinning disk. In it, he laments the continued use of “ancient storage technology” and foresees “a post-storage world” where the industry is “devoid of the painfully outdated yet ubiquitous spinning disk.”
His take fits our view that hard disks are headed the way of the floppy though he says that in some scenarios, HDDs will be used like the way tape is used today. I suppose the alternative is simply recycling.
Venezia says “rethinking centralized storage is a necessary part of this transition [to all SSD].” He is right, but you also have to rethink how you evaluate SSD storage solutions too. Innovation is making hardware dumb and software smart. SSD hardware in the enterprise is fast becoming a commodity and prices are falling in line with that movement. More important than the media itself when comparing your SSD options, software capabilities should trump hardware. Focusing too much on hardware and not enough on the software platform wrapped around it introduces greater risk of quick obsolescence and the inability to adapt to changing requirements.
Read the rest of this entry »
Tags:application performance, data protection, Database Performance, Flash, Flash SSD, I/O bottlenecks, I/O wait, IOPS, K2, Kaminario, Kaminario K2, latency, MLC, Storage Performance
Posted in SSD TCO | No Comments »
Posted March 6th, 2013
|
Flash Immortal
Discovery could extend Flash life from 10,000 write cycles to more than 100 million
By Kaminario
|
An interesting Gizmodo article entitled Hot Flash: The Secret To SSDs That Will Outlast The Universe discusses a new discovery claimed by Taiwanese Flash manufacturer Macronix that by periodically applying intense heat to MLC Flash for milliseconds they can increase its life from 10,000 write cycles to more than 100 million. How does it work? Flash writes erode transistor insulation slowly until eventually the chip is no longer viable. Add a quick jolt of 800 degree-centigrade heat and you can actually heal the insulation for another 5 to 10,000 writes.
No doubt this is an important discovery, but the truth is that the life limitations of flash are greatly exaggerated even today. Flash manufacturers and solution vendors have come up with a raft of technologies, including wear leveling, advanced error correction, and sophisticated digital signal processing, to ensure your Flash device can last almost as long or longer than a typical hard disk.
Read the rest of this entry »
Tags:application performance, Database Performance, I/O bottlenecks, IOPS, latency, MLC, OLTP, reads, Storage Performance, writes
Posted in SSD TCO | No Comments »
Posted February 26th, 2013
|
Measure Performance Over Time and with Three Dimensions
ANALYST SURVEY LOOKS AT REAL-WORLD IOPS REQUIREMENTS
By Kaminario
|
Tom Coughlin and Jim Handy, respected storage analysts, conducted a survey on how many IOPS are really needed for several popular enterprise applications to better understand real-world IOPS requirements. They asked respondents about their critical storage applications and how many IOPS those applications needed. More than 80 percent of their initial respondents said that their key app was a database, an OLTP solution or a Cloud storage service. About half of the respondents said their IOPS needs were in the 1k to 100k range. Coughlin and Handy note that other factors like latency contribute to overall performance, not just speed. You can view research details from a presentation that Tom did at the SNIA Storage Developer conference, How Many IOPS is Enough? or get the full report at How Many IOPS Do You Really Need?
Measure Consistent Performance
I am bringing this up now as a way to comment on how Kaminario looks at performance. We do not see performance as a point in time such as timing a runner in a single forty-yard dash. True performance is measured by consistency of speed and results over time and as capacity and applications are added. You shouldn’t necessarily buy an all-Flash array based of the performance of a single app because there are so many variables that can come into play down the road. This is one of the reasons that Kaminario is taking the general purpose storage approach with the K2. We think it is smarter to evaluate a Flash storage system in the context of your entire storage environment and future app needs. You don’t want to buy a system that gives you great value up front but falls short in mixed workload environments. You don’t want to have to buy a separate storage system when you introduce a new application or need to drastically scale up capacity.
Read the rest of this entry »
Tags:application performance, Database Performance, Flash SSD, high availability, I/O bottlenecks, IOPS, K2, Kaminario, latency, RDBMS Performance, Solid State Storage, Storage Performance
Posted in SSD Storage Performance | No Comments »
Posted December 5th, 2012
|
What Kind of SSD Solution is Best for your Needs?
STORAGE SWITZERLAND POST SUGGESTS QUESTIONS TO ASK
By Kaminario
|
Judging from industry chatter and published articles, it is clear that the debate about all SSD versus a hybrid SSD/HDD array is picking up. Our blog touched on this a little bit recently. This strikes me as a good thing. While we believe the industry is on a journey to the time when HDDs will be used like tape is today, naturally there are waypoints. People need to balance their storage needs and budget with all available options to them.
This brings me to an article I recently saw from Storage Switzerland titled Flash SSD May be the Answer, but It Raises Many Questions. It highlights some of the questions storage managers should start asking their vendors about SSD storage to arrive at the best type of SSD solution for their needs.
The story suggests questions about sharing capabilities, high availability and a few other topics that organizations should ask vendors. When we speak with customers, we try to ask questions of our own. So here are three questions we always ask customers considering a hybrid or all-SSD solution:
Read the rest of this entry »
Tags:application performance, Database Performance, DataProtect, Flash SSD, I/O Performance, I/O wait, IOPS, Kaminario K2, latency, MLC, OLTP, RDBMS Performance, Solid State Storage, Storage Efficiency, Storage Performance
Posted in SSD Solutions | No Comments »
Posted July 30th, 2012
 |
Kaminario Sits Atop Storage Performance Mountain with Record-Setting SPC-1 Result
SOLID STATE SAN STORAGE PERFORMANCE GETTING MORE AFFORDABLE
By Gareth Taube, Vice President Marketing, Kaminario
|
Kaminario today announced independent audit results of the K2’s price performance conducted by the Storage Performance Council. In summary, we blew the doors off the SPC-1 Results, breaking the SPC-1 performance record as the first certified results exceeding one million IOPS, while also cutting the cost-per SPC-1 IOP by more than half. These results are the best in the history of the SPC-1. In addition, the K2 sustained its IOPS levels in a 24-hour SPC-1 test run. 
Sources:
Kaminario K2: http://www.storageperformance.org/results/benchmark_results_spc1#a00118
IBM System Storage SAN Volume Controller 6.2 with IBM Storwize® V7000 Disk Storage: http://www.storageperformance.org/results/benchmark_results_spc1#a00113
HP P10000 3PAR V800 Storage System: http://www.storageperformance.org/results/benchmark_results_spc1#a00109
Texas Memory Systems RamSan-630: http://www.storageperformance.org/results/benchmark_results_spc1#a00105
Validation for SPEAR —As I see it, this is a big validation for our Scale-out Performance Storage Architecture (SPEAR). SPEAR distinguishes us from others in the industry with its unique ability to orchestrate mixed workloads over a distributed cluster of solid-state storage — all within an environment that supports advanced data protection features. Plus, as a software innovation, SPEAR offers substantial flexibility to adapt to evolving needs. All this leads to better application performance typically between 200 and 2,000 percent.
A testament to the robustness of the SPEAR design is that it sustained the 1.2M SPC-1 IOPS performance over 24 hours, maintaining performance well beyond the 8-hour period achieved by previous SPC competitive tests. This proves that the SPEAR design is a stable and consistent architecture, something that many of our customers who run continuous analytical processing applications can appreciate.
What It Means —High-end solid-state SAN storage performance is getting increasingly affordable. Clearly, solid-state price/performance ratios are getting better and better over similar HDD ratios. Kaminario’s SPC-1 results are not necessarily a total tipping point for SSD, but when you more than double the performance record and cut the cost-per-IOP record by more than half to $0.40 per SPC-1 IOP, it is a significant milestone.
Read the full SPC-1 report at: http://www.storageperformance.org/results/benchmark_results_spc1/#a00118
Kaminario’s continued focus on improving performance at lower cost within a full data protection environment is aimed at mainstream database and OLTP applications in the enterprise data center. Growing application sizes and complexity means that high-performance storage requirements are no longer for niche applications only.
Last January, we proclaimed 2012 the Year of the SSD. Given our outstanding, unsurpassed SPC-1 Result, I can’t help but think that the remainder of this year and all of next year will be even bigger for solid-state storage.
Tags:application performance, Database Performance, DRAM, DRAM Storage, DRAM-based SSD Appliances, I/O bottlenecks, IOPS, K2, Kaminario, latency, OLTP, SPC-1
Posted in SSD Storage Performance | No Comments »
Posted June 18th, 2012
|
SSD Array Architectures and Latency: The Proof is in the Pudding
PROOF THAT SSD ARCHITECTURE MATTERS
By Gareth Taube, Vice President Marketing, Kaminario
|
There are some interesting discussions revolving around Robin Harris’s StorageMojo blog—too bad he’s taken a vacation just when the discussion’s getting really good. The topic is SSD architectures and whether disk-form-factor SSD arrays can possibly provide the kind of performance speed-hungry applications require.
At Kaminario we’ve long argued that a disk-form-factor SSD architecture is inherently flawed. While the SSD itself might deliver on performance, inevitably the legacy storage controllers and other legacy components become a performance and reliability bottleneck, increasing latency and reducing throughput as several SSD’s vie for the same limited pipe.
Read the rest of this entry »
Tags:architecture, benchmarks, bottleneck, disk form factor array, Flash, Howard Marks, K2, Kaminario, latency, Robin Harris, SAS, SATA, SSD, SSD array, SSD performance, storage controller, StorageMojo
Posted in SSD Architectures | No Comments »
Posted May 22nd, 2012
 |
Go Beyond Raw Numbers When Considering All Flash Arrays
EVALUATION REQUIRES LOOKING AT IOPS & LATENCY IN REALISTIC SCENARIOS
By Eyal Markovich
|
Last week, I participated in an interesting conversation about performance measurement. The jist of it was that IOPS and latency are related. Properly evaluating an array’s performance requires examination of both. For example, what happens to latency when IOPS double?
Latency is the critical performance measurement for most applications. This is especially true for relational databases because they often drive a few thousands of IOPS. Improve latency and you will improve application performance. Keep in mind latency changes based on workloads.
Below are a few things to consider when comparing performance numbers of all SSD SAN storage.
Look Beyond Raw Numbers in Realistic Workloads
If you are discussing performance with a Flash array vendor, look beyond an IOPS number and a latency number. Ask how they work together in realistic workloads. It is possible to achieve a stellar latency number if you run a small, synthetic IO transaction. The question is what happens when you test a workload that really stresses the array? Or, what is the performance on a realistic workload such as a database with 70 percent reads and 30 percent writes with difference IO sizes? Unlike a card in a server which usually serves only one application, an array will serve several. Each application needs improved latency. However, the more applications you host on an array, the higher the IOPS. What is the effect of adding IOPS to latency by each application? You should work with a SSD vendor to understand the correlation between IOPS and latency on demanding workloads.
Read the rest of this entry »
Tags:Flash, Flash SSD, I/O wait, IOPS, K2, Kaminario, latency, Storage Performance
Posted in Data Protection | No Comments »
Posted May 7th, 2012
|
Array Rumble at Storage Tech Field Day
SSD ARRAY VENDORS FACE OFF OVER SOLID-STATE STORAGE ARCHITECTURES
By Gareth Taube, Vice President Marketing, Kaminario
|
Another interesting seminar at Stephen Foskett’s Tech Field Day, hosted by Nigel Poulton, addressed the best architecture for an SSD array. Participants included Thomas Isakovich, CEO and Founder of Nimbus Data Systems, Umesh Maheswhari, CTO and founder of Nimble Storage, Jonathan Goldick, software CTO at Violin Memory, and Dave Wright, founder and CEO of SolidFire. It was a pretty lively debate, with Goldick grinning broadly through much of it and taking jabs at the others. One couldn’t help but wonder what he was grinning about.
The truth is, there was a lot of disagreement over the best array architecture and sometimes the argument got a tad heated and personal, much to the delight of the audience. However, there were three things everyone could agree to. First, the best architecture is one that provides an ideal balance of scalability, share ability, reliability, and performance, not performance alone. Second, for all but the few most performance- and latency-sensitive applications, it’s more important to provide consistent, predictable performance for an array of applications, than to provide the absolute best performance. And third, the best architecture is a mix of commodity hardware and a software architecture designed from the ground up for SSD. Sound familiar? Read the rest of this entry »
Tags:architecture, cache, Dave Wright, deduplication, disk, Flash, hybrid, Jonathan Goldick, Kaminario, latency, Nigel Poulton, Nimble Storage, Nimbus Data Systems, off-the-shelf hardware, scale-out architecture, SolidFire, SSD, SSD array, Stephen Foskett, Tech Field Day, Thomas Esakovich, tiering, Umesh Maheshwari, Violin Memory Systems
Posted in SSD Storage Performance | 2 Comments »
Posted April 8th, 2012
|
Murphy’s Law and Home-Grown SSD HA
BAKED IN BEATS WRAP AROUND
By Gareth Taube, Vice President Marketing, Kaminario
|
In his recent StorageSearch blog, High Availability Enterprise SSD’s, Zsolt Kerekes discusses why building a do-it-yourself HA solution typical of data center Fibre Channel or IP SAN hard disk array installations is not a viable option for an enterprise SSD array. The reasons, according to Zsolt, boil down to performance, flexibility of use, risk, complexity, and scalability. I would add Murphy’s Law.
Zsolt points out that any home-grown HA solution sitting in front of an SSD storage controller is likely to introduce considerable latency and time to the recovery process. When it comes to the mission critical applications typically running on SSD arrays, such as online transaction processing, time really is money and simply not something you want to sacrifice. “Wrap around” HA, as Zsolt calls it, also introduces architectural complexity and controller configuration issues that can gum up the works.
Read the rest of this entry »
Tags:cluster, data center, data corruption, DataProtect, enterprise SSD array, Fibre Channel, high availability, IP SAN, K2, Kaminario, latency, Murphy's Law, N+1, Online Transaction Processing, outage, SSD, Storage, storage node, StorageSearch, Zsolt Kerekes
Posted in SSD Storage Performance | No Comments »
