The cluster we tuned is hosted on AWS and is comprised of 6 hi1.4xlarge EC2 instances, with 2 1TB SSDs raided together in a raid 0 configuration. The cluster’s dataset is growing steadily. At the time of this writing, our dataset is 341GB, up from less than 200GB a few months ago, and is growing by 2-3GB per day. The workload on this cluster is very read heavy, with quorum reads making up 99% of all operations.Some careful GC tuning here. Probably not applicable to anyone else, but good approach in general.
this is definitely one to send a consultation document response to
I’ve seen a lot of hand-wringing from techies in San Francisco and Silicon Valley saying “Why are we so hated?” now that there’s been a more vocal contingent of people being critical of their lack of civic responsibility. Is it true that corruption and NIMBYism have kept affordable housing from being built? Sure. Is it true that members of the tech industry do contribute tax dollars to the city? Absolutely. But does that mean techies have done enough? Nope.
Some basic succinct data structures. […] The main highlights are: a novel, broadword-based implementation of rank/select queries for up to 264 bits that is highly competitive with known 32-bit implementations on 64-bit architectures (additional space required is 25% for ranking and 12.5%-37.5% for selection); several Java structures using the Elias–Fano representation of monotone sequences for storing pointers, variable-length bit arrays, etc. Java code implementing minimal perfect hashing using around 2.68 bits per element (also using some broadword ideas); a few Java implementations of monotone minimal perfect hashing. Sux is free software distributed under the GNU Lesser General Public License.
Sugar blocks concrete from setting. This I did not know