> >> IO connections are the 2 domains where PC can't be compare with > >> mainframe.
And never will. > It look like my idea :) But 8 PCI-E link is 2.4 GB/s in each direction... > That's why saturate a 300 MB/s is soon enough. (random access of HD did > not reach 40 MB/s) Disk access isn't always random access. Sequentially reading a large file on a decent filesystem (e.g. FFS) doesn't do much seeking. I just got 49194139 bytes/sec reading a large file (~10GB) on FreeBSD. Linux on the same machine could only manage 35741200 bytes/sec. The biggest disk speed problem today for SATA is the lack of NCQ support. Developers are having trouble getting info on how to access NCQ for various controllers. Thus, this is an area where an open-source SATA controller could help significantly. > > - SAN/NAS storage standalone controller(iSCSI/FC/Infiniband) for SAN/NAS > > manufacturers and DIY people. This might be useful. I'm thinking Gigabit Ethernet to a bunch of SATA ports. SATA port multipliers are another possibility. The last time I looked I only found 1 5-port PM in the market. > > Real-time intrusion detection Who is the target market for intrusion detection on a NAS box? Put the thing in a locked closet / machine room. > > or virus > > detection would be feasible with an additional FPGA card. Is the open source crowd interested in virus detection? Don't run virus server -> problem 99.9999999999999% solved. > > - Solid state disk Not to rain on your parade, but... These exist in the marketplace. It isn't obvious to me that open source would add a huge amount of value here. (do other projects first) And I don't see a huge market for these. Some mainboards can have a huge amount of main memory. _______________________________________________ Open-graphics mailing list [email protected] http://lists.duskglow.com/mailman/listinfo/open-graphics List service provided by Duskglow Consulting, LLC (www.duskglow.com)
