Garantia Data Unveils Automated In-Memory NoSQL Cloud ServiceJuly 19, 2012 No Comments
SOURCE: IDV News
Garantia Data has launched the beta release of what the company says is the first fully-automated, in-memory NoSQL cloud service that offers reliable Memcached and infinitely-scalable Redis data store systems.
Facebook, Twitter, Instagram and Pinterest already use Memcached and Redis to support high-performance and rapid growth. That said, both data store technologies could use improvement, according to Garantia Data CEO Ofer Bengal. Memcached can be not that reliable and Redis doesn’t scale very well, and both need constant operational care, Bengal said.
“We have leveraged sophisticated technology to solve real industry pains,. Our new in-memory NoSQL cloud reinvents the way people use Memcached and Redis,” he said in a statement.
Garantia Data’s approach aims to provide customers fully-automated scalability in a completely “hassle-free” cloud service. The company offers Memcached users full reliability with no data loss and it offers Redis users infinite scalability without compromising on any of the Redis commands, he said.
In specific, Garantia Data looks to fill in these gaps with its dynamic-auto-sharding technology. The approach virtualizes multiple cloud servers into an infinite pool of memory, this allows datasets to scale autonomously and continuously to sizes in the terabytes and even petabytes, according to Bengal. Further, because Garantia Data’s technology doesn’t need to be constantly managed, devs don’t need to deal with nodes, clusters, server lists, configuration, scaling and failure recovery. This also avoids the loss of data, he added.
The Garantia “shrads” are one of the main keys to the architecture. The company explans how a shard works this way:
The shards created autonomously inflate, deflate, multiply or reduce according to the dataset’s size and the measured performance of each shard. This means dataset can autonomously and continuously grow from a few megabytes to terabytes and even petabytes. A dataset is constantly replicated, so if a node fails, an auto-switchover mechanism guarantees data is served without interruption. The technology supports various data persistence options without compromising on data integrity or throughput and latency. One can also back up a dataset to a remote persistent storage for disaster recovery purposes.
The Garantia Data in-memory NoSQL cloud is currently available on Amazon Web Services free of charge to early adopters during the beta phase. When it is released to GA (generally availability) later this year, the company will offer the solution as a “pay as you go” subscription model, so customers will only pay for the actual memory that they use, according to Bengal. The company intends to expand its offering to other public clouds later this year, he added.
One analyst gave the new technology a thumbs up.
“Every web-scale application developer that uses distributed memory caching should take a close look at this break-through solution from Garantia Data,” said Paul Burns, president of cloud computing industry analyst firm Neovise, in a statement.
Garantia Data’s solution provide Memcached and Redis capabilities as a cloud service through an API, it also makes them reliable and scalable, Burns noted. The technology also eliminates the time consuming, error-prone administrative processes usually involved in establishing and maintaining private caching deployments, he added.
The Garantia Data Memcached Cloud was built from the ground up over a true Memcached cluster architecture and with a highly reliable infrastructure. The solution will distribute a dataset across multiple shards in multiple nodes of the company’s Memcached cluster. The service will also provide constant monitoring of all data shards to ensure optimal performance. As needed, more shards and nodes can be added.
Garantia Data also provides Memcached cluster replication, data persistence, backup and auto-failover capabilities. The cluster also fully-automates all operational activities, including launching or taking down nodes, upgrading software, scaling, clustering and recovering from failures.CLOUD COMPUTING