Download the authoritative guide: Enterprise Data Storage 2018: Optimizing Your Storage Infrastructure
It is another year and another try at predicting the future in our complex storage industry. The year in review for 2010 can be summarized in two words -- stagnation and consolidation: stagnation as there has been limited innovation, which is totally understandable given the economic downturn, and consolidation as Tier-2 and Tier-3 storage players have been acquired right and left as the large vendors prepare for the economic upturn.
So how did I do with my predictions for 2010?
I said that FCoE will become available end-to-end and major vendors will support it. At least two major vendors support FCoE end-to-end, but cost reductions have not been realized yet. Given the economy, many companies are sticking with upgrading their current Fibre Channel architecture rather than deploying completely new Fibre Channel over Ethernet (FCoE) architecture.
Score: 1 (but not a strong 1)
I predicted that PCIe 3.0 would make it to market this year and that you would be able to buy PCIe 3.0 motherboards. The PCIe 3.0 standard was not released until this month, and some vendors are just showing sample motherboards at trade shows.
I predicted that server vendors will address the end-to-end data integrity problem Wrong again. This is not expected until next year. It keeps getting put off year after year.
I said that the flash SSD market will undergo consolidation and that STEC will get competition in the enterprise flash space. STEC does have competition now, but there has been limited consolidation in the market.
Score: 0 for consolidation
Score: 1 for competition
I said that SSD flash usage will increase among RAID vendors, and that controller cards will provide better support.
I predicted that 10GbE will become the standard for high-end PCs. This is true industry wide.
I stated that NFSv4.1 will enter the market in 2010. This did not happen, although almost all of the NFSv4.1 framework is in the latest Linux kernel, anda number of vendors are demonstrating products at trade shows.
I stated that 40Gbps and 100Gbps Ethernet will continue to march on the way to product development. In June of 2010 the standards were completed, and today products have been demonstrated at shows such as Supercomputing in November.
Last but not least, I made a prediction about clouds, stating: Clouds are good for some things, but just like the storage service providers and application service providers in the late 1990s, or grid computing in the early part of this decade, clouds will meet a similar fate. SSPs and ASPs are providing services for some applications, but they are not going to solve all problems for all enterprises, as there is just not enough network bandwidth, the latencies for some applications are too high, and the security problem end-to-end has not be solved in a standard way (see Why Cloud Storage Use Could Be Limited in Enterprises). Clouds have not solved all problems, there are no standards, and yet the hype continues. It reminds me of the 1980s commercial: Wheres the beef?
My final tally was not very good compared to other years: six correct and four wrong, which is really bad. My only excuse is that I thought that the economy would have been better, creating more need for innovation. As I mentioned in my last article, the economy is what is causing new disk technology to be delayed (see Budgeting for Disk Storage, 2010 2013). And disk drives are just one area where there has not been significant investment.
Now for my ten storage predictions for 2011.
There will not be huge changes next year because there has not been a great deal of investment in technology over the last two to three years. Some of the predictions below are things I expected to happen much sooner when the economy was better, and with a slight glimmer of economic improvement I am hoping these technologies will finally make it into the mainstream.
1. Declustered RAID will become mainstream
There are currently a few vendors that support declustered RAID, but with the movement to 4TB drives next year and increasingly larger and larger storage environments this will become a standard across the industry. Of course, it is a buyer beware situation, and you can expect an article from me on this topic next year.
2. Large RAID vendors will support FCoE targets
I hoped this would have come sooner, but it cannot be too far from realization in the market. FCoE and Ethernet will dominate the market and Fibre Channel will be relegated to legacy environments.
3. PCIe 3.0 hits the market, but no 16-lane slots or cards
PCIe 3.0 doubles performance over PCIe 2.0 per lane, and today even mid-range motherboards have PCIe 2.0 16-lane slots. In 2011 we will not see performance per lane double for PCIe 3.0 but we will just have 8-lane slots on motherboards and 8-lane cards. Given that CPU and memory bandwidth performance has significantly increased, the lack of PCIe bus bandwidth increase in 2011 will impact I/O performance, and what might be hurt most is PCIe-based SSDs.
4. SSDs will achieve 2,000,000 IOPS performance
PCIe-based SSDs recently hit the one million IOPS mark and it is likely that two million IOPS will be possible next year.
5. SAS RAID/JBOD cards will hit 1,000,000 IOPS
SAS-based JBOD cards are making huge performance strides and some RAID cards now have performance that it as good as many external RAID devices. The one million IOPS number is a big jump, but I think it will happen in late 2011.
6. pNFS products will finally be released
The pNFS framework has finally been released in the Linux kernel and vendors are starting to create pNFS layout drivers for NAS devices and file systems. 2011 should be the year of pNFS product releases from a number of vendors.
7. Silent data corruption in file systems will be addressed in hardware
T10 DIF/PI will finally make it to market. End-to-end implementation (from host to disk) has been more difficult than expected. Of course, some file systems such as ZFS and PAN-FS have this functionality already, but having it in hardware has some significant advantages.
8. The dramatic declines in tape unit sales will significantly slow
The decline in the tape market is well known, but I believe the decline has stopped. The number of PB of tape shipped will actually increase in 2011 and tape unit sales could exceed 2010 levels.
9. SSD density growth will slow
Both Jeff Layton and I have written about the expected slowdown in SSD density growth (see Why Solid State Drives Won't Replace Spinning Disk and Why Flash Drive Density Will Stop Growing Next Year).
I think the current highest-capacity 2.5-inch SAS SSD is around 500GB. This will not grow more than 40% over the next year, which is significantly down from previous years.
10. 2.5-inch drives will surpass 3.5-inch drives in sales in Q4 2011
The disk drive market is changing, and 2.5-inch drives are the future for IOPS performance, watts per GB and other reasons. I had expected 2.5-inch drives to dominate much sooner, but the economy caused RAID vendors to delay developing new backend drive trays for 2.5-inch drives.
Happy Holidays and best wishes for a more prosperous 2011!
Henry Newman, CEO and CTO of Instrumental, Inc., and a regular Enterprise Storage Forum contributor, is an industry consultant with 29 years experience in high-performance computing and storage.
Follow Enterprise Storage Forum on Twitter.