- PalDB : embeddable write-once key-value store written in Java by LinkedIn crowd. This solution kind of sit between json/file and level/rocksdb [github]
- RDMA + HTM : very nice concept, would love to see the extension of HTM to NVM tech in order to close the loop. T His would deliver tremendous speed up capability. But also another type of concurrency model to understand.
- Metrics 2.0 : emerging set of conventions, standards and concepts around time series metrics metadata
Wednesday, October 28, 2015
Tuesday, October 27, 2015
Links of the day 27/10/2015 : #Swarm vs #FLeet vs #Kubernetes vs #Mesos 50 years of #datascience, #google temporal graph DB
- Swarm vs FLeet vs Kubernetes vs Mesos : well balanced article but it always end up being a question of affinity to a certain philosophy and how "experimental" (or cool depending of your ecosystem) . [other tech discussed here]
- 50 years of Data Science : pretty much the must read document if you want to understand what datascience is , its evolution and key tech behind it.
- Badwolf : temporal graph store using the RDF model (triplet) by google
Friday, October 23, 2015
- Openstack at scale at CERN : 5k hypoerisor, 15k VM, up to 1 Millions vm, the scale of the CERN openstack deployment is really what is driving the effort to push openstack beyond its limits.
- HEPIX fall 2015 : Hepix is a really interesting workshop as it provide real use case report and experience of large scale IT infrastructure and as time evolved pretty much HPC cloud centric solution.
- Performance comparison of KVM and Docker instances in OpenStack : well no surprise here, docker faster all the way
Wednesday, October 21, 2015
Links of the day 21/10/2015: #blockchain consensus protocol, Awesome #aws resources, Killing patents
- Awesome AWS : A curated list of awesome AWS libraries, open source repos, guides, blogs, and other resources with a popularity rating system built in :)
- Blockchain Consensus Protocols : a little bit of blockchain beyond fintech (but still a lot of fintech in this slidedeck).
- So You Want To Murder a Software Patent: talk by Jason Scott on how to deal with annoying patents
Tuesday, October 20, 2015
- Velocity 2015 : speakers slides , lots of very good talks, particularly the sampling strategy for large scale monitoring and breaking stuff to make them more reliable.
- Cloud HPC guide : when you want your workload doesn't fit the enterprise definition.
- HPC networking trends : while the trend is always for bigger network the interesting aspect is the focus also on improving the MTTI as it currently takes days to recover from failure.
Thursday, October 15, 2015
- EPaxos, transactions and the next 700 Paxos systems : well distributed consensus will always be a highly discussed topic and wheel will be reinvented many time because not everybody like the size of the.
- Energy-Efficient In-Memory Data Stores onHybrid Memory Hierarchies : propose code instrumentation in order to identify the best object placement for performance/power efficiency. Nice idea however lack of practicality for production environment. Also it assume a pretty stable behavior of the system.
- IISWC-2015 : 2015 IEEE International Symposium on Workload Characterization
Wednesday, October 14, 2015
- Streaming video 10Gb and beyond : well seems everybody is bypassing the kernel nowadays.
- Building Scalable Stateful Services : Age old cycle - state - no state - etc..
- Benchmark comparisons of erasure codes : pick carefully , but not all are equal. Again there is no one size fit all.
Tuesday, October 13, 2015
- Intel cache partitioning : why Intel added cache partitioning, the feature is really useful in order to avoid cache pollution and noisy neighbor effect in virtual environment. Until there is an automated way to provision the configuration (autonuma style maybe). It will remain niche. Too bad this is badly needed as the number of cores to cache ratio keep increasing.
- Irreversible Failures : Lessons from the AWS DynamoDB Outage, and why you should always be prepared for white/gray/black swan events. Make your system ductile not brittle..
- SOPS 2015 : interesting pick from microsoft research : ironfleet distributed system validation, and distributed transaction. From Intel we have the virtual cpu validation
Monday, October 12, 2015
- VISC : update on the virtual core technology by soft machine. Interesting concept, however like the Itanium, really hard to pull off as it requires a specific combination of software and hardware to work. Lets see if they fare better than Intel in that domain.
- Queue splitting: want to use DPDK for some traffic and the rest go through the standard stack? checkout the bifurcated drivers!
- NVDIMM and persistent memory in QEMU : cool talk about persistent memory tech for virtualization. This open up a whole new realm of possibility.
Wednesday, October 07, 2015
- Consensus in the wild : lot of way to achieve consensus out there but which one is the most efficient for your requirements. Answer : it depends and we have no consensus.
- IronFleet: Tool for proving corectness of practical distributed systems using TLA-style state-machine refinement and Hoare-logic verification.
- AtlasDB : Transactional Distributed Database layer on top of K/V store.
Tuesday, October 06, 2015
Links of the day 06/10/2015 : the long road to quantum computing , collaboration tool, and cloud scheduling
- ORCHESTRA : open source system to orchestrate teams of experts and machines on complex projects.
- Expensive Road Ahead For Quantum Computing : well nobody said it was going to be easy.
- Cloud Computing Resource Scheduling and a Survey of Its Evolutionary Methods : A good insight in evolutionary scheduling method for cloud or other similar problem.
Friday, October 02, 2015
- Linaro Connect : For those of you using ARM based system. [all the presentations][videos]
- Nomad : cluster manager and scheduler designed for microservices and batch workloads by hashicorp.
- Kudu : another storage solution for hadoop.