Cloudian, Splunk Unchain Compute, Storage
Splunk has decoupled the computing and storage layers of its architecture, providing a way for customers to scale storage independent from computing.
The machine data analytics specialist working with storage partner Cloudian has cleaved the two layers with Splunk’s (NASDAQ: SPLK) new smart storage feature combined with Cloudian’s HyperStore remote storage. The combination allows independent scaling of computing and Splunk storage as the volume of machine and unstructured data soars.
Splunk’s indexer feature is designed to reduce dependence on expensive local storage, shifting the bulk of data to Cloudian’s remote storage. The combination is said to yield total-cost-ownership savings as high as 60 percent when compared with indexer-based approaches while boosting storage scaling, Splunk claims.
Those savings are derived by replacing traditional data tiering approaches in which companies buy new storage for each technology upgrade, shifting data from hot, warm and cold tiers. “This approach has proven very difficult to scale as data volumes have skyrocketed,” the partners noted.
As data transitions from hot to warm, the Splunk indexer uploads a copy of the intermediate, or warm,” data bucket to Cloudian’s remote storage while retaining a bucket locally on the indexer. Those buckets are rotated out as active data sets are added. Local storage is then used for “hot” buckets and cached copies of “warm” buckets that might contain, for example, data required for future searches.
Storage integration with Cloudian uses Amazon Web Services’ (NASDAQ: AMZN) Simple Storage Service (S3) object storage to separate computing and storage layers. The combination facilitates on-premises petabyte storage without adding an indexer. That storage pool is also separated from a Splunk indexer used to review documents and compile indexers used in data searches.
Along with Amazon S3, Cloudian also said its multi-threaded storage architecture integrates with other public cloud storage services that can be used either to replicate or tier data.
The partners said the new storage option is available either as a preconfigured appliance or as software running either on bare metal servers or virtual machines. Appliance storage capacities run as high as 1.5 petabytes.
Along with Splunk, Cloudian also provides storage options for data management applications from Pure Storage, Quantum, Rubrik, Veeam and others.