List of Scalaris Releases

• Scalaris 0.9.0 (codename “Vriesea scalaris”) - February 29, 2016 - scalaris-0.9.0.tar.gz
• Scalaris 0.8.2 - July 24, 2015 - scalaris-0.8.2.tar.gz
• Scalaris 0.8.1 - July 12, 2015 - scalaris-0.8.1.tar.gz
• Scalaris 0.8.0 (codename “Picoides scalaris”) - July 10, 2015 - scalaris-0.8.0.tar.gz
• Scalaris 0.7.2 - October 23, 2014 - scalaris-0.7.2.tar.gz
• Scalaris 0.7.1 - September 30, 2014 - scalaris-0.7.1.tar.gz
• Scalaris 0.7.0 (codename “Stauroderus scalaris”) - April 28, 2014 - scalaris-0.7.0.tar.gz
• Scalaris 0.6.1 - October 11, 2013 - scalaris-0.6.1.tar.gz
• Scalaris 0.6.0 (codename “Conus scalaris”) - August 19, 2013 - scalaris-0.6.0.tar.gz
• Scalaris 0.5.0 (codename “Saperda scalaris”) - October 11, 2012 - scalaris-0.5.0.tar.gz
• Scalaris 0.4.1 - March 22, 2012 - scalaris-0.4.1.tar.gz
• Scalaris 0.4.0 (codename “Pomacea scalaris”) - January 25, 2012 - scalaris-0.4.0.tar.gz
• Scalaris 0.3.0 (codename “Rhinechis Scalaris”) - July 15, 2011 - scalaris-0.3.0.tar.gz

Scalaris 0.9.0 (codename “Vriesea scalaris”) - February 29, 2016 - scalaris-0.9.0.tar.gz

(partly supported by the EU project IES Cities http://iescities.eu/)

Packaging:

• add a libscalaris-dev[el] package with the C++ API (rpm, deb)
• Maven: let Maven also create a source jar file for the Java-API
• Maven: allow proper snapshot versions for the maven repository at https://scalaris-team.github.io/scalaris/maven

API:

• CPP-API: provide a new C++ client library
• DataNucleus: provide a new storage back-end for Scalaris
• add a Scalaris adapter for YCSB
• new functions in all APIs for generic routing table operations

Business Logic:

• rrepair: considerably improve the set reconciliation algorithms’ performance and run more processes in parallel
• rrepair: further reduce the reconciliation costs
• rrepair: re-use left-over failure probability from phase 1 for phase 2
• rrepair: increase accuracy in several formulaes of the set reconciliation algorithms
• rrepair: support creating multiple Merkle trees for fewer message rounds (rr_merkle_num_trees configuration variable)
• rrepair: remove support for aligning bit sizes (no improvements there)
• transactions: speed up write ops when the transaction log already contains a read op on the same key
• RM: optimise the traffic of state exchanges during ring maintenance
• RT: improve routing parallelism by handling more messages in the routing process itself instead of inside the dht_node process
• RT: improve performance along the routing path
• support for dynamic replication factors given by the replication_factor configuration variable
• pid_groups: shuffle pids in find_a/1 for more better distributed (parallel) requests

Infrastructure:

• pid_groups: group names are now atoms instead of strings (allows faster access)
• scalarisctl: do not set any default Erlang scheduler flags (allow changes through the ERL_SCHED_FLAGS environment variable)
• extend the example scripts for running Scalaris on a SLURM cluster
• add SLURM scripts for running Basho Bench
• use configure to select the routing table type (--with-rt=rt_chord)
• change hanoidb detection to configure --enable-hanoidb=<dir>
• change erlang_js detection to configure --enable-erlang-js=<dir>
• dht_node_monitor: switch messages on/off via configuration variable dht_node_monitor
• msg_delay: remove extraneous error messages for trigger messages of killed processes
• prbr: allow write_filter to return value to qwrite calling process
• update yaws to version 2.0.2
• support for Erlang R14B04 up to 18.2.3 and current otp master

Tests:

• allow running single test groups via make <test suite>:<group name>_GROUP

Tools:

• gen_component: allow processes to provide their own start functions
• trace_mpath: write whole message of log_info to the tex output

Bugs:

• fix cleanup of some special cases with transactions
• fix several math issues and increase the math functions’ performance
• fix some errors which occurred in large scale deployments
• fix several warnings reported by Coverity Scan
• fix numerous more bugs

Scalaris 0.8.2 - July 24, 2015 - scalaris-0.8.2.tar.gz

(partly supported by the EU project IES Cities http://iescities.eu/)

Business Logic:

• rrepair: minor performance improvements

Bugs:

• rrepair: fix a rare crash due to hash collisions
• rrepair: fix calculation of the next MaxItemCount in the merkle tree repair protocol (was one level behind)
• tests: fix rare crashes in some unit tests
• fix a few more minor bugs

Scalaris 0.8.1 - July 12, 2015 - scalaris-0.8.1.tar.gz

(partly supported by the EU project IES Cities http://iescities.eu/)

Business Logic:

• rrepair: allow specifying the minimum number of hash bits to use to create non-approximate algorithms

Bugs:

• rrepair: fix a badarith if rr_recon_version_bits is set to ‘variable’
• rrepair: fix a bug in the accuracy of the evaluation
• fix two crashes in gossip

Scalaris 0.8.0 (codename “Picoides scalaris”) - July 10, 2015 - scalaris-0.8.0.tar.gz

(partly supported by the EU project IES Cities http://iescities.eu/ and the EIT ICT Labs project MCData)

Packaging:

• also install docs on Arch Linux
• fix lintian errors and warnings for Debian-based packages
• fix permissions on the log directory (owned by the ‘scalaris’ user) and the config files (not owned by the ‘scalaris’ user anymore!)
• adapt packages for newer distributions

API:

• Java-API: integrate new OtpErlang library (1.5.12 from Erlang 17.4)
• Python-API: do not include a shebang with a pre-defined path to python anymore
• Python-API: fix unclosed sockets in various tests
• Python-API: increase python{2,3} test speed
• Ruby-API: try to fallback to the normal json module if the gem is missing
• Ruby-API: considerably increase the ruby test speed
• remove the Publish/Subscribe API

Business Logic:

• rrepair: faster and more accurate bloom filter repair protocol
• rrepair: faster, more accurate and less bandwidth-consuming merkle tree repair protocol
• rrepair: reduce memory footprint of merkle_tree
• rrepair: more accurate and less bandwidth-consuming trivial repair protocol
• rrepair: add a new simple hash (‘shash’) algorithm
• rrepair: switch to (the more efficient) push-only resolve phase for synchronisation
• rrepair: add replica repair algorithm evaluation scripts
• RM: faster cyclon cache integration, especially during node jumps
• comm_layer: envelops can be nested
• comm_layer: allow send options for comm:send_local/3 and msg_delay (currently only ?quiet supported)

Infrastructure:

• FD: simplify the API, support envelopes and remove own cookie handling
• FD: extend the failure detector with generic notifications (not just crashes) to subscribed nodes
• DB: add an (experimental) mnesia back-end
• DB: add an (experimental) hanoidb back-end
• DB: make the back-end configurable via the scalaris.cfg option ‘db_backend’
• improve the speed of several utility functions
• scalarisctl: add ‘-l’ parameter for specifying the logdir
• scalarisctl: do not use the ‘-s’ parameter anymore, instead specify the start type with the new ‘-t’ parameter
• scalarisctl: allow starting several dht_nodes per vm with given keys via the new parameter ‘-j’
• update yaws to version 1.99
• make “$HOME/.scalaris/log” the default log dir
• support for Erlang R14B04 up to 18.0.2 and current otp master
• add example scripts for running Scalaris on a SLURM cluster
• move sources to github: https://github.com/scalaris-team/scalaris
• new project homepage: http://scalaris.zib.de

Tests:

• reduce log spam in proto_sched tests and only print proto_sched results when the tests fail
• extended and improved various test suites

Tools:

• proto_sched: support callbacks
• proto_sched: add support for user msg shepherds
• proto_sched: better error handling when sending messages to dead (local) processes
• tester: add value creator and type checker for orddict:orddict()
• tester: full support for Erlang 18.0
• let ‘make dialyzer’ run over test as well and ignore some false-positives for Erlang >= 18.0 using the ‘-dialyzer()’ attribute

Bugs:

• slide: fix some race conditions with ‘jump’ operations
• JSON-API: fix test_and_set not working correctly inside req_list
• fix numerous more bugs

Scalaris 0.7.2 - October 23, 2014 - scalaris-0.7.2.tar.gz

(partly supported by the EU project IES Cities http://iescities.eu/ and the EIT ICT Labs project MCData)

Packaging

• fix ArchLinux packages with newest Java versions

Demonstrator “Wiki on Scalaris”

• fix the separate count key optimisation not using Scalaris’ increment operation

Business Logic

• rrepair: let the trivial algorithm assume the worst case in order to always meet the configured “recon probability of one error” (p1e)
• rrepair: fix the trivial algorithm having an effectively doubled p1e
• rrepair: fix the bloom algorithm having an effectively tripled p1e
• rrepair: allow disabling byte-alignment

Bugs

• fix a few minor bugs

Scalaris 0.7.1 - September 30, 2014 - scalaris-0.7.1.tar.gz

(partly supported by the EU project IES Cities http://iescities.eu/ and the EIT ICT Labs project MCData)

Packaging

• add support for new distribution versions
• support systemd and SELinux
• include daemon for monitoring Scalaris through JMX

API

• Java-API: integrate new OtpErlang library (1.5.10 from Erlang 17.3)

Demonstrator “Wiki on Scalaris”

• fix storing template back-links for magic words
• fix ARTICLE_COUNT partitioning with hashes

Business Logic

• more robust (still experimental) support for active load balancing with Karger and Ruhl’s algorithm including more flexible “load” definitions
• rm_tman: less overhead by only sending unknown nodes to neighbours
• slide: better support for slide aborts
• node join: less overhead during joins, especially in setups with huge lists of known nodes
• cyclon: move to new gossip framework
• vivaldi: move to new gossip framework
• monitor: move performance monitor to the basic_services group (once per VM)

Infrastructure

• several smaller performance optimisations
• support for Erlang R13B01 up to 17.3 and current otp master

Tests

• some new unit tests and higher test coverage

Documentation

• add rrepair sequence diagrams

Bugs

• fix numerous bugs

Scalaris 0.7.0 (codename “Stauroderus scalaris”) - April 28, 2014 - scalaris-0.7.0.tar.gz

API

• Java-API: integrate new OtpErlang library (1.5.9 from Erlang 17.0)

Demonstrator “Wiki on Scalaris”

(supported by 4CaaSt http://iescities.eu/ and Contrail http://contrail-project.eu):

• add support for reading 7z dumps

Business Logic

(partly supported by the EU project IES Cities http://iescities.eu/ and the EIT ICT Labs project MCData):

• add an experimental Map-Reduce framework on top of Scalaris
• add experimental support for active load balancing with Karger and Ruhl’s algorithm including more flexible “load” definitions
• rrepair: completely new merkle tree sync protocol with an order of magnitude lower traffic costs (use dynamic signature sizes based on a “recon probability of one error” (p1e) in the leaf nodes)
• rrepair: add a trivial probabilistic reconciliation protocol using the “recon probability of one error” (p1e)
• rrepair: replace bloom_fpr parameter with a generic “recon probability of one error” (p1e)
• rrepair: add a trivial reconciliation phase for the differences identified by bloom
• rrepair: more efficient resolve also including a list of keys to request
• gossip: completely new (more flexible) gossip framework
• slide: add support for ‘jump’ operations
• slide: do not create a timer for each received message (use a periodic cleanup)
• comm_layer: tune TCP connection parameters to increase throughput and latency (no delay_send)
• comm_layer: close idle TCP connections, support for no_keep_alive connections (used by cyclon, for example)
• comm_layer: try to bundle more messages
• node join: less overhead during joins, especially in small rings
• rm_tman: less overhead by not adding nodes in the cache to the dn_cache
• rt_chord: skip trying to contact nodes in the own range
• rt_chord: stop stabilize when the own node is reached

Infrastructure

• intervals: more compact (and transfer-friendly) representation
• trigger: replace modules with msg_delay:send_trigger/2
• config: replace implementation with a public ets table
• tune the garbage collection of some core processes and periodically garbage collect all processes
• reduce the number of generated atoms and the number of ets tables used
• several smaller performance optimisations
• support for Erlang R13B01 up to 17.0 and current otp master

Tests

• improve test coverage (manual tests and random tests)
• change test definitions and allow groups of tests, e.g. test (default), test-skipped, all_TESTS, all_with_cover_TESTS, performance_TESTS, proto_sched_TESTS, type_check_TESTS

Documentation

• add rrepair documentation

Tools

• proto_sched: new API, better determinism, more strict self-control
• proto_sched: support for short-lived processes
• proto_sched: more information in get_infos
• trace_mpath: more flexible LaTeX exports with more detailed message info
• trace_mpath: support meaningful PID names when tracing remote messages
• tester: full support for Erlang 17.0
• make: re-compile after Emakefile changes

Bugs

• rrepair: fix unnecessary feedback for KVV items of the same version
• bulkowner: some fixes for (gracefully) leaving nodes
• fix numerous more (less severe) bugs

Scalaris 0.6.1 - October 11, 2013 - scalaris-0.6.1.tar.gz

Packaging

• add ArchLinux ruby API packages

API

• Java-API: add Maven build support
• Java-API: add CircularByteArrayOutputStream#clear()
• Java-API: fix ConnectionPool#getConnection(timeout) throwing IllegalMonitorStateException if a single ConnectionPool is used by multiple threads and no more connections are available
• Java-API: add a “-monitor” command line parameter
• api_monitor: return the latency and stddev values of the micro-benchmark executed by monitor_perf for node and service performance

Demonstrator “Wiki on Scalaris”

(supported by 4CaaSt http://www.4caast.eu/ and Contrail http://contrail-project.eu):

• separate list counters from their list partitions for a better data layout
• fix high memory use of the Scalaris import if the import is slow
• use tomcat 7.0.42

Business Logic

• rrepair: reduce overhead of ART reconciliation
• rrepair: allow resolving of multiple merkle node leaves with a single resolve request
• rrepair: don’t create resolve requests for empty intervals
• rrepair: reduce overhead of update_key_entry requests (use a single request with all the data instead of one request for each item)
• rrepair: allow arbitrary intervals in interval_upd and interval_upd_send resolve requests again
• rrepair: when hashing merkle_tree/ART children, also include the represented interval (fixes indistinguishable empty leaf nodes in ART)
• tx_tm_rtm: re-enable takeover by rtms on tm crash
• rt_chord: only re-build the RT if the pred or succ processes change or the own new node ID is not between the new pred and succ any more

Infrastructure

• DB: improve performance of fold[lr] implementations by a factor of 2
• FD: get rid of annoying, wrongly raised log warnings
• RM: provide a more generic Reason for RM subscriptions
• RM: remove trigger infections
• RT: don’t trigger an update when a slide finishes
• RT, FRT: reduce the number of messages sent on lookups
• log: don’t exit with a badmatch if our error_logger is not the only one

Tests

• add protocol scheduler tests for slide, join and leave

Documentation

• user-dev-guide: add section on scalarisctl checkinstallation
• user-dev-guide: clarify the section about how to set up Scalaris

Tools

• proto scheduler: continue in case of send errors
• top: improve process messages output

Bugs

• node join: fix not being able to join a system with passive load balancer if the number of items in the DB is too high
• tx_tm_rtm: fix wrong asserts
• fix node resposibility check not always including message forward and db_range intervals
• tx_tp: add missing snapshot number in tp_do_commit_abort message
• bulkowner: respect the nodes’ DB ranges and forward requests for non-responsible ranges
• dn_cache: fix reporting wrong PIDs back as zombies after node reboots
• fix some more (less severe) bugs

Scalaris 0.6.0 (codename “Conus scalaris”) - August 19, 2013 - scalaris-0.6.0.tar.gz

Packaging

• add ArchLinux packages
• add support for new distribution versions

API

• no more timeouts in client APIs
• Java-API: re-worked the request and result list handling -> move result processing to the operation classes
• Java-API: better support for custom operations
• Java-API: support the new partial reads: ReadRandomFromListOp and ReadSublistOp
• Java-API: compile with “vars” debug info
• Java-API: integrate new OtpErlang library (1.5.8 from R16B) with fixed support for compressed binaries
• Java-API: add back-ports from the Wiki on Scalaris demonstrator: ** list-change operations: ScalarisChangeListOp and ScalarisListAppendRemoveOp ** MultiMap classes are now in de.zib.tools ** CircularByteArrayOutputStream
• Java-API: fix hostname issues with Erlang and Java
• Java-API: slightly changed the delete API
• JSON-API: add API for auto-scale requests
• Python-API: add API for auto-scale requests
• Python-API: use default socket timeout
• Ruby-API: use default socket timeout
• all APIs: support lists of composite types

Demonstrator “Wiki on Scalaris”

(supported by 4CaaSt http://www.4caast.eu/ and Contrail http://contrail-project.eu):

• allow monitoring via JMX in the FourCaastMonitoringPlugin
• support for getting random articles via the new partial read op
• new optimisation scheme “Buckets with Write Cache” - uses a single big list
• to read from and the rest of the buckets to write to
• improve import and dump-processing (faster, more memory-efficient)
• add on-the-fly conversion to the different optimisation schemes during import
• (only one prepared DB dump needed now)
• several UI enhancements and rendering fixes
• update bliki lib (includes code ported to upstream)
• add auto-import ability
• use tomcat 7.0.33

Business Logic

• replace common message tags with integers to reduce bandwidth
• more flexible read operations (easier to extend)
• add support for the following partial reads: random_from_list and sublist
• save bandwidth by not returning the full value for write operations (only the version is required)
• new DB back-end implementation with a smaller and cleaner interface
• faster DB get_chunk processing
• tx: allow overwriting old/outdated DB entries
• tx: allow overwriting old/outdated write-locked entries
• tx: allow setting write lock on old/outdated read-locked entries
• tx: always reply when the majority replied during read
• tx: make sure that if not_found is reported to the user (while reading), a write cannot go through if it is not also based on not_found
• tx: committing a test_and_set op on a non-existing entry now fails as well (the op itself already returned the failure)
• tx: add a 2s delay to wait for slow learner_decide answers before cleaning up (results in a faster state cleanup after the fourth response)
• tx: small performance improvements in several modules
• rm: only add alive, non-leaving nodes
• rm: if a predecessor crashes, start repairing the range (rrepair)
• rrepair: stabilised rrepair (not considered experimental any more)
• rrepair: also update entries with existing but outdated WriteLocks
• rrepair: several performance improvements (bloom, merkle_tree, art and rrepair processes in general)
• rrepair: re-design of rr_recon
• rrepair: don’t offload heavy work onto the dht_node (increases responsiveness of the dht_node process during replica repair)
• rrepair: improve db_generator tool and random_bias binomial distribution used for tests
• rrepair: support differently configured nodes (use the same reconciliation structure parameters)
• rrepair: de-activate self-repair (a node with multiple copies of the same items does not need a reconciliation structure to repair some of them)
• rrepair: activate rrepair periodically every 10 minutes with a probability of 33%
• slide v2.0: fewer message to initiate a slide
• slide v2.0: generic (asynchronous) call-backs for different ring maintenance algorithms
• slide v2.0: re-work handling of planned next operations, e.g. used by incremental slides
• slide v2.0: don’t directly work on the DB any more (there may be more data needed to slide) - let dht_node_state decide
• slide v2.0: activate incremental join and leave operations
• slide v2.0: actively report graceful node shutdown to the local FD of the leaving node to inform subscribers
• slide v2.0: code clean-up
• slide v2.0: some fixes for incremental slides
• slide v2.0: more robust in general
• more smooth node joins by also reporting when a join is not possible due to a running slide at the existing node
• passive load balancing: random selection of (equally qualified) nodes
• add new routing algorithms FRT-Chord (flexible routing tables) and GFRT-Chord (supports proximity routing and data centers) as alternatives to Chord (see rt_frtchord and rt_gfrtchord modules)
• add auto-scale framework, e.g. for cloud environments (supported by Contrail http://contrail-project.eu/) which is able to scale the deployment to maintain a given target latency of executed transactions
• cache config reads in the process dictionary for better performance
• cyclon: if the cache is empty, try one of the nodes in known_hosts
• add support for consistent snapshots (experimental)

Infrastructure

• add a daemon to monitor Scalaris via JMX
• disable message compression (only client values are compressed - the rest is too expensive, at least on GbE)
• support for distributions with python3 available as “python” and python2 as “python2”
• support for Ruby 1.9
• yaws 1.96 (with patch to compile on otp master and a patch to fix a performance regression)
• support for Erlang R13B01 up to R16B01 and current otp master

Tests

• add test suite to find memory leaks
• let “make test” run the major test suites and “make test-skipped” for some more (time-consuming) tests
• clean-up ring after timetrap timeout failures via common test hook
• new ?compare macro for custom comparison functions
• higher test coverage with more random-testing via the “tester”

Documentation

• user-dev-guide: add user tutorial on using scalaris
• user-dev-guide: add a section about the slide protocol
• user-dev-guide: extended description of scientific background
• add replica repair sequence diagrams
• better code descriptions

Tools

• gen_component: synchronous breakpoint set and delete for more deterministic usage
• trace_mpath: allow selective tracing via filter fun
• trace_mpath: fix several triggers becoming infected by trace_mpath resulting in infinite tracing
• trace_mpath: improve latex output of traces
• tester: copy dictionary to worker threads
• tester: add support for more types, e.g. neg_integer(), gb_rees
• tester: better type check error reporting
• tester: print tester last calls when aborting unit tests (timeout or exception)
• tester: add support for constraints in type specs (“when is_subtype(A,B)”)
• web debug interface: add cluster graph visualisation
• web debug interface: display vivaldi distance
• web debug interface: add IP addresses and ports to the ring charts and tables
• web debug interface: allow navigating to the web interfaces of shown nodes
• top: support for showing messages in message queue of an inspected PID
• top: support for showing larger dictionary values
• allow recursive reply_as envelopes
• experimental protocol scheduler to check protocols with random message interleavings (see proto_sched module)

Bugs

• fix RM handling of (out-dated) nodes with the same ID as newly added nodes
• fix ganglia integration not working any more
• restore the ability to start nodes at a specific key via }
• fix some memory leaks in the tx system
• fix statistics of comm_connection (not send in some cases, not overflow-aware)
• use /bin/bash instead of /bin/sh which may not result in a bash session
• fix init.d scripts not checking for existing processes correctly
• fix dc_clustering
• fix numerous other bugs

Scalaris 0.5.0 (codename “Saperda scalaris”) - October 11, 2012 - scalaris-0.5.0.tar.gz

Packaging

• new init.d script to start Scalaris
• added chef scripts to deploy Scalaris nodes
• improved Windows start scripts (support for R15B01 and R15B02, don’t close command prompt window immediately after shutdown if double-clicked)
• more flexible scalarisctl (arbitrary parameter order, allow setting cookie, ports and number of nodes in VM via parameters, allow using screen for daemonised sessions, allow graceful leave via “gstop” command, new “status” command)
• support for new linux distributions (Fedora 17, Ubuntu 12.04, openSUSE 12.2)
• let scalarisctl checkinstallation also perform runtime tests for the APIs

API

• allow Scalaris monitoring via JMX through the Java API
• added an executor-service to the Java-API (de.zib.scalaris.executor)
• added a node discovery daemon to the Java-API
• allow compressed communication between the Java-API and Erlang for increased performance, especially if the two are on separate nodes
• added VM management support to the JSON- and Python-API
• added transaction log filtering to the Java-API, i.e. only sent the needed parts of the tlog back to Erlang and re-combine the result
• fixed api_tx:req_list_commit_each/1 not running requests in parallel ** do not assure any order of requests, even if on same key!

Demonstrator “Wiki on Scalaris”

(supported by 4CaaSt http://www.4caast.eu/):

• allow different partitioned data models for better performance and scalability
• allow logging of user requests
• added support for checking whether another article exists (approximate) ** show link colours based on this check
• added check for bad page titles
• allow SERVERNAME and SERVERPATH in config for setups with load balancers
• reduced memory footprint of Wiki data in Scalaris
• support for newer wiki xml dumps
• added support for using a MediaWiki-like SQLite-DB backend for e.g. filtering
• improved overall performance
• several rendering fixes

Business Logic

• added (experimental) support for replica repair (disabled by default) (thanks to Maik Lange)
• added monitoring of memory statistics (also available via web interface)
• better error reporting in the failure detector
• reduced message overhead by UIDs and message/tuple tags
• reduced overall message size of transactions: ** do not include the (uncompressed) value in messages of the read phase of write operations ** do not include the value in init_TP messages
• allow VM-flag “first” to be set via config file
• gather overall connection statistics in comm_stats (also available via web interface)
• reduced erroneous failure messages on node shutdown
• integrated comm_layer into comm_server
• better scalability in pid_groups (find processes round-robin in find_a/1)
• several changes to improve overall performance and/or CPU time at the nodes

Tests

• support for more types in the runtime type-checker
• verify several API functions via runtime type-checker (also test private functions if possible!)

Tools

• distributed protocol visualisation via trace_mpath (text-based or latex-file for graphical presentation)
• better profiling via top for Erlang processes
• better debugging, e.g. via ASCII supervisor-tree rendering in verbose mode

Bugs

• fixed memory leaks in read and write operations
• fixed memory leaks in tx_tm_rtm
• prevent potential endless loops in tx_tm_rtm
• fixed inform RTMs sometimes informing the wrong RTMs
• fixed numerous other bugs

Scalaris 0.4.1 - March 22, 2012 - scalaris-0.4.1.tar.gz

Packaging

• new official ConPaaS packages (http://www.conpaas.eu/)
• install rubygem dependencies in Debian postinstall scripts for Ruby API
• improved Windows start scripts (if set, uses the ERLANG_HOME environment variable to find Erlang, otherwise searches for Erlang in common paths)

Bugs

• better tx cleanup (should fix rare occurance of duplicate client inform)
• forward additional parameters of the start scripts to new syntax of scalarisctl

Scalaris 0.4.0 (codename “Pomacea scalaris”) - January 25, 2012 - scalaris-0.4.0.tar.gz

API

• new functions for incremental data change: ** test_and_set: check for a provided old value before setting a new one ** add_on_nr: increment a numeric value ** add_del_on_list: append or delete entries from a list value
• added VM API to manage Scalaris nodes inside an Erlang virtual machine
• added monitoring API to retrieve some live metrics
• added a connection pool convenience class (Java, Python)

Demonstrator “Wiki on Scalaris”

(supported by 4CaaSt http://www.4caast.eu/)

• improved performance of page edits
• improved performance of Wikipedia dump loading
• several rendering fixes

Business Logic

• improved handling of large values by reducing overhead of transaction log handling (empty TLog after commit), no copy of value in TLog returned to user after read requests)
• eliminated timeouts in data hand-over protocol (relies on fd now)
• added a DB subscribe mechanism, e.g. to become informed when locks are freed
• fixed a strong consistency issue in the tx protocol
• gather some run-time statistics and expose them via the APIs and the web debug interface

Infrastructure

• support for Erlang 15B
• fd now also uses feedback from TCP layer
• made message sending more flexible (gets an option list)
• added and corrected several Erlang type specifications
• added scripts to create Scalaris images for OpenNebula
• added tools for using Scalaris as the Database as a Service component in ConPaaS (http://www.conpaas.eu/) which is part of the EU project Contrail (http://contrail-project.eu/)
• added a separate communication channel for priority messages, e.g. fd (reduces falsely reported node crashes under heavy load)

Tests

• added runtime type-checker for random testing extended unittests

Documentation

• updated documentation to extended APIs

Bugs

• fixed numerous bugs

Scalaris 0.3.0 (codename “Rhinechis Scalaris”) - July 15, 2011 - scalaris-0.3.0.tar.gz

API

• new API with interoperable bindings to Java, Python, Ruby, and JSON
• support for several data types, including strings, integers, JSON objects, binary objects.
• new transaction interface with support for bundled requests for better latency.
• separate APIs to access the raw DHT, a DHT with replication, and the transactional DHT

Demonstrator

• added Wikipedia-hosting using Scalaris as demonstrator application

Business Logic

• fault-tolerant startup: start Scalaris when a quorum of the known_hosts becomes available (option -q in bin/scalarisctl)
• perform data hand-over when nodes join/gracefully leave (also works when transactions are executed concurrently)
• added passive load balancing (when a node joins a ring, it samples several other nodes and joins at the node that balances the number of stored items the most)
• completely rewritten transaction layer (more modular, more extendible, less latency)
• modularized / reimplemented Paxos algorithm, so the algorithm can also be used outside transactions (e.g. used for quorum-startup)
• switched almost all components to our component framework ‘gen_component’
• added gossiping for estimating e.g. the number of nodes or the average load in a ring
• more reliable unreliable look-up
• better ring start-up on slow networks

Infrastructure

• Vivaldi and topology inference
• support for Erlang 13B01 and newer
• faster TCP/IP communication between Scalaris nodes
• completely rewritten failure detector framework for more accurate detection of node failures
• added numerous Erlang type specifications
• extended unittests

Tests

• added own random testing framework that reads type specifications and scans the source code for constants to generate proper random test-data
• extended gen_component with breakpoint-support for debugging and testing (perform deterministic pseudo-random message interleaving tests)
• added numerous unittests
• added language-binding interoperability tests

Documentation

• extended, but - as always - by far not enough…

Bugs

• fixed countless bugs