As you may have noticed I have been silent for a while. The reason for that is that we have been concentrating on our 3.0 release which carries quite a significant number of new features and improvements to our product. One of the main changes in product distribution will be that we now are going to have two editions: GridGain Community Edition and GridGain Enterprise Edition. Community edition still remains feature-rich and we are as always committed in providing stellar forum-based support for it. However, certain features in the 3.0 product line had taken significant amount of effort and we are simply not going to be able supporting them for free.So in addition to Current Feature List, here is the list of additional features coming in GridGain 3.0 product line (as you see, we have been busy ;-)
GridGain 3.0 Community Edition
Distributed Cache
- Local and Fully-Replicated caches.
- Zero-Deployment - cached objects and classes are automatically deployed to all grid nodes.
- Full Replication and Invalidation modes.
- Distributed transactions including Optimistic and Pessimistic modes.
- One-Phase and Two-Phase commit.
- Feather-weight Eventually Consistent transactions.
- Distributed locking.
- Cache data disk-overflow using file-based GridSwapSpaceSPI implementation.
- Read-Through and Write-Through modes with pluggable persistent storage.
- Pluggable expiration policies including LRU, LFU, FIFO, or user-specified.
- Synchronous and Asynchronous cache operations, including asynchronous transaction commit.
- Cache metrics and cache event notifications.
- Synchronous and Asynchronous cache warm-up from other grid nodes on start-up.
- Cache warm-up from persistent storage on start-up.
- Full integration with GridGain MapReduce implementation.
- Integration with Hibernate Level2 cache.
- One of the richest Cache API on the market including support for functional programming with closures and predicates.
- Node discovery based on list of initial IPs.
- Amazon S3-based automatic node discovery.
- Support for Continuous Mapping (even after initial split).
- Task scheduling.
- Ability to swap large objects to disk for tasks with large memory consumption.
- Many other bug fixes and enhancements.
GridGain 3.0 Enterprise Edition
(in addition to all features from community edition)
Partitioned Cache
- Partitioned Cache based on consistent hashing key-to-node assignment.
- Zero-Deployment - cached objects and classes are automatically deployed to all grid nodes.
- Integration with GridGain MapReduce for collocation of computations and data.
- Pluggable consistent hashing algorithms.
- Cache data disk overflow based on H2 database with support for data encryption.
- Distributed query support based on standard SQL syntax.
- Distributed query support for combination of in-memory and on-disk data
- Custom indexes for queried data with optional overflow to disk for large indexes (including data encryption for overflown indexes).
- Optional data affinity replicas for redundancy.
- Load-balancing or data and computation requests among replicas.
- Group based discovery with support for extremely large and partially connected clusters.
- Support for no-connectivity and one-way connectivity between groups.
- Heterogeneous support for joining different clouds and private data centers into one coherent grid.
- Merge support for disconnected clouds.
- Java-based consistent API for controlling any cloud (including Amazon EC2).
- Ability to control multiple clouds at the same time.
- Configurable strategies for adding and removing cloud resources based on current system characteristics.
- Seemless support for starting/stopping grid nodes on the cloud.
- Data encryption for disk swap space for task with large memory consumption.
- Custom shell for grid monitoring and management.
- Ability to script any grid operation.
- Simple per-per-use model - if grid node is not used, you don't pay.
View comments