The average consumer might not know much — if anything — about high-end technologies like pNFS, PCIe and 10Gb Ethernet (10GbE), but they could revolutionize the consumer PC market in ways that most consumers can’t even begin to imagine.
High-end technologies tend to move down market over time — along with their pricing — and I believe these will be no different. In this case, these new technologies have the potential to address an age-old problem for the home user that is getting worse over time.
PC Upgrades and Data Migration Problems
One of the biggest problems for home PC users is migration of data to new systems. Most consumers get new computers every three to five years. During that period, the amount of data they amass from cameras, video, music and everything else increases fairly dramatically. Getting that data from one system to another is no easy task. You can use USB, Ethernet or some other method, but it can be a slow process.
A cottage industry of products and services has sprung up just to move your data, including external storage devices that are basically low-end NAS devices. Data Robotics (Drobo), HP (NYSE: HPQ), EMC (NYSE: EMC), Seagate (NASDAQ: STX), Western Digital (NYSE: WDC), and many others, have all stepped in to fill the need. I am not going to talk about the merits of products or methods, but needless to say, some products are designed for some parts of the market better than others, and some are easier to use and to migrate to new technologies than others.
So you have small NAS boxes getting more affordable and disk drive costs dropping significantly, but the performance to move data to a new system is constant or at best a bit faster, so with so much more data, it is a big problem to reliably move it all. Another thing to remember is that these home NAS devices are relatively slow compared to the speed of a hard drive. The performance available today for home PCs ranges in the marketing literature from 100 MB/sec to 66 MB/sec. We all know what marketing speak is, but for the sake of this discussion, I will use the upper end at 100 MB/sec. The fastest current NAS network connectivity is 1Gbit Ethernet, but plenty of PCs are limited to Firewire or USB-2 speeds.
Using 1GbE hardware, NAS speeds are limited to a peak of 125 MB/sec. The real number when you add NFS or CIFS communications protocols and file system overhead is more like 30 MB/sec for write and maybe 50 MB/sec for read. So at best this current crop of NAS boxes is much slower than disk drives, between one-half and one-third the performance of current disk drive technology. This all changes when 10GbE becomes available later this year at commodity pricing; then the channel becomes much faster than the drives (see Falling 10GbE Prices Spell Doom for Fibre Channel). For now, though the usage model for these low-end NAS devices is for bulk storage. They are not fast enough to replace local disk drives yet and the latency is high, but with 10GbE, pNFS and faster, lower-latency PCIe buses, the game will change.
When a home user wants to upgrade storage, generally what happens is users get new storage with a new machine and have to move the files on the old system to the new one. This is a great business model for many organizations that support national chains that charge good money to move the files to the new system. With technologies like Drobo and others that allow for storage migration inside the box to higher density storage, data migration becomes significantly easier, as you just have to add new disk drives and wait for the box to rebuild itself. The problem today is these technologies are too slow for many home users doing video editing or editing large photos or temporary files; the write performance is just too slow and read is not much better. These performance problems get eliminated with the commoditization of 10GbE. The Ethernet connection will be faster than the drives, which has never happened to my memory. 10GbE speeds are around 400 MB/sec to 600 MB/sec for some vendors, including the NFS overhead. Imagine that for video editing.
A Home PC Revolution Coming
The end results of all this, I think, is that people will begin to want to separate storage from their operating systems at home, which is no different than what is happening in enterprises or small sites. The problem is that at home, the storage performance needed to be in the host, at least until now. Now storage and the operating system can be separated for the home user. And given the falling cost of flash, my next home PC may be a 128GB flash drive for the OS, applications and temporary files, and a NAS bulk storage device connected with 10GbE. You might need more or less flash, but your PC will run cooler and can be upgraded without the need for a new system and file migration, plus your boot time will be much faster. You can likely get away with less memory, as swap space is so much faster swapping to a hard drive. In short, we may begin to see the enterprise storage and server model move into the home.
The one problem in all this is laptops; mobile users, of course, have different data access needs. But perhaps faster network speeds and data syncing software and services could revolutionize the laptop model too, as is already happening with smartphones. Data could become centralized for everything, which would have tremendous benefits for our increasingly mobile society.
I am looking forward to the day when this type of configuration becomes available at a reasonable cost, and I believe that is likely not too far down the road. All we need is 10GbE on the motherboard and in low-end NAS devices, and that could occur in the next year or so. That would allow much easier (and cheaper) upgrades to systems and storage. Of course, there will need to be changes to operating systems (Windows and Mac OS) to simplify management. And many online backup environments do not support NAS, so something has to change to allow NAS devices to be backed up. This type of environment is still too complicated for the masses without some support both from the NAS and OS sides of the equation, but I think it will happen out of necessity. And once you have data separated from the OS and applications, technologies like deduplication will move down market to home PCs.
The end result for enterprise storage could be to accelerate the move to unified storage and converged networks. Selling large numbers of equipment to the commodity world in NAS format would emphasize the development of NAS end-to-end, from the home to the enterprise, and would likely result in additional features and functions (and ease of use) that could be deployed from the home to the enterprise. It could also reduce the amount of development in the SAN world and increase the disparity between NAS and SAN for features and support, speeding up the merger of the SAN and NAS worlds. All this may not be as far off as you think.
Henry Newman, a regular Enterprise Storage Forum contributor, is an industry consultant with 28 years experience in high-performance computing and storage.
See more articles by Henry Newman.
Follow Enterprise Storage Forum on Twitter