Friday, 23 April 2010
4.1.0. ALPHA3 is out
I’ve just cut Infinispan 4.1.0.ALPHA3, codenamed Radegast. This release contains a number of fixes and bugs reported in 4.0.0 Starobrno as well as earlier alphas, and is quite likely to be the last alpha before a feature-complete 4.1.0.BETA1 is released.
A detailed changelog is available. The release is downloadable in the usual place.
If you use Maven, please note, we now use the new JBoss Nexus-based Maven repository. The Maven coordinates for Infinispan are still the same (groud id org.infinispan, artifact id infinispan-core, etc) but the repository you need to point to has changed. Setting up your Maven settings.xml is described here.
Enjoy! Manik
Tags: release alpha radegast
Tuesday, 13 April 2010
Boston, are you ready for Infinispan?
The JBoss World/Red Hat Summit organisers have asked me to put together a short note on what to expect at the conference with regards to Infinispan, and this has been published on the conference website. I thought I’d share this with you as well.
In addition to the conference, this is also the first time JBoss is running JUDCon (JBoss Users and Developers Conference), a free event (limited on a first-come first-served basis) for the community, also in Boston. Be sure to sign up as places are limited!
Look forward to seeing you there! Manik
Tags: conference judcon jbossworld
Tuesday, 06 April 2010
Infinispan 4.1Alpha2 is out!
We’ve just released Infinispan 4.1.0.Alpha2 with even more new functionality for the community to play with. Over the past few weeks we’ve been going backwards and forwards in the Infinispan development list discussing Infinispan’s binary client server protocol called Hot Rod and in 4.1.0.Alpha2 we’re proud to present the first versions of the Hot Rod server and java client implementations. Please visit this wiki to find out how to use Hot Rod’s java client client and server. Please note that certain functionality such as clients receiving topology and hashing information has not yet been implemented.
Besides, Infinispan 4.1.0.Alpha2 is the first release to feature the new LIRS eviction policy and the new eviction design that batches updates, which in combination should provide users with more efficient and accurate eviction functionality.
Another cool feature added in this release is GridFileSystem: a new, experimental API that exposes an Infinispan-backed data grid as a file system. Specifically, the API works as an extension to the JDK’s File, InputStream and OutputStream classes. You can read more on GridFileSystem here.
Finally, you can find the API docs for 4.1.0.Alpha2 here and again, please consider this an unstable release that is meant to gather feedback on the Hot Rod client/server modules and the new eviction design.
Cheers, Galder & Mircea
Tags: hotrod server
Tuesday, 30 March 2010
Infinispan eviction, batching updates and LIRS
DataContainer abstraction represents the heart of Infinispan. It is a container structure where actual cache data resides. Every put, remove, get and other invoked cache operations eventually end up in the data container. Therefore, it is of utmost importance the data container is implemented in a way that does not impede overall system throughput. Also recall that the data container’s memory footprint can not grow indefinitely because we would eventually run out of memory; we have to periodically evict certain entries from the data container according to a chosen eviction algorithm.
LRU eviction algorithm, although simple and easy to understand, under performs in cases of weak access locality (one time access entries are not timely replaced, entries to be accessed soonest are unfortunately replaced, and so on). Recently, a new eviction algorithm - LIRS has gathered a lot of attention because it addresses weak access locality shortcomings of LRU yet it retains LRU’s simplicity.
However, no matter what eviction algorithm is utilized, if eviction is not implemented in a scalable, low lock contention approach, it can seriously degrade overall system performance. In order to do any meaningful selection of entries for eviction we have to lock data container until appropriate eviction entries are selected. Having such a lock protected data container in turn causes high lock contention offsetting any eviction precision gained by sophisticated eviction algorithms. In order to get superior throughput while retaining high eviction precision we need both low lock contention data container and high precision eviction algorithm implementation – a seemingly impossible feat.
Instead of making a trade-off between the high precision eviction algorithm and the low lock contention there is a third approach: we keep lock protected data container but we amortize locking cost through batching updates. The basic idea is to wrap any eviction algorithm with a framework that keeps track of cache access per thread (i.e. ThreadLocal) in a simple queue. For each cache hit associated with a thread, the access is recorded in the thread’s queue. If thread’s queue is full or the number of accesses recorded in the queue reaches a certain pre-determined threshold, we acquire a lock and then execute operations defined by the eviction algorithm - once for all the accesses in the queue. A thread is allowed to access many cache items without requesting a lock to run the eviction replacement algorithm, or without paying the lock acquisition cost. We fully exploit a non-blocking lock APIs like tryLock. As you recall tryLock makes an attempt to get the lock and if the lock is currently held by another thread, it fails without blocking its caller thread. Although tryLock is cheap it is not used for every cache access for obvious reasons but rather on certain pre-determined thresholds. In case when thread’s queue is totally full a lock must be explicitly requested. Therefore, using batching updates approach we significantly lower the cost of lock contention, streamline access to locked structures and retain the precision of eviction algorithm such as LIRS. The key insight is that batching the updates on the eviction algorithm doesn’t materially affect the accuracy of the algorithm.
How are these ideas implemented in Infinispan? We introduced BoundedConcurrentHashMap class based on Doug Lea’s ConcurrentHashMap. BoundedConcurrentHashMap hashes entries based on their keys into lock protected segments. Instead of recording entries accessed per thread we record them in a lock free queue on a segment level. The main reason not to use ThreadLocal is that we could potentially have hundreds of threads hitting the data container, some of them very short lived thus possibly never reaching batching thresholds. When predetermined thresholds are reached eviction algorithms is executed on a segment level. Would running eviction algorithm on a segment level rather than entire data container impact overall eviction precision? In our performance tests we have not found any evidence of that.
Infinispan’s eviction algorithm is specified using strategy attribute of eviction XML element. In addition to old eviction approaches, starting with release 4.1.ALPHA2, you can now select LIRS eviction algorithm. LRU remains the default. Also note that starting with 4.1ALPHA2 release there are two distinct approaches to actually evict entries from the cache: piggyback and the default approach using a dedicated EvictionManager thread. Piggyback eviction thread policy, as it name implies, does eviction by piggybacking on user threads that are hitting the data container. Dedicated EvictionManager thread is unchanged from the previous release and it remains the default option. In order to support these two eviction thread policies a new eviction attribute threadPolicy has been added to eviction element of Infinispan configuration schema.
Does eviction redesign based on batching updates promise to live up to its expectations? Ding et al, authors of the original batching proposal, found that their framework increased throughput nearly twofold in comparison with unmodified eviction in postgreSQL 8.2.3. We do not have any numbers to share yet, however, initial testing of BoundedConcurrentHashMap were indeed promising. One of our partner companies replaced their crucial caching component with BoundedConcurrentHashMap and realized a 54% performance improvement on the Berlin SPARQL benchmark for their flagship product. Stay tuned for more updates.
Cheers,
Vladimir
Tags: eviction concurrency data structures
Friday, 12 March 2010
No time to rest, 4.1.0.Alpha1 is here!
"Release quick, release often", that’s one of our mottos at Infinispan. Barely a couple of weeks after releasing Infinispan 4.0.0.Final, here comes 4.1.0.Alpha1 with new goodies. The main star for this release is the new server module implementing Memcached’s text protocol.
This new module enables you to use Infinispan as a replacement for any of your Memcached servers with the added bonus that Infinispan’s Memcached server module allows you to start several instances forming a cluster so that they replicate, invalidate or distribute data between these instances, a feature not present in default Memcached implementation.
On top of the clustering capabilities Infinispan memcached server module gets in-built eviction, cache store support, JMX/Jopr monitoring etc… for free.
To get started, first download Infinispan 4.1.0.Alpha1. Then, go to "Using Infinispan Memcached Server" wiki and follow the instructions there. If you’re interested in finding out how to set up multiple Infinispan memcached servers in a cluster, head to "Talking To Infinispan Memcached Servers From Non-Java Clients" wiki where you’ll also find out how to access our Memcached implementation from non-Java clients.
Finally, you can find the API docs for 4.1.0.Alpha 1 here and note that this is an unstable release that is meant to gather feedback on the Memcached server module as early as possible.
Cheers, Galder
Tags: release memcached