Page 266 - Big data - Concept and application for telecommunications
P. 266
5 Big data - Concept and application for telecommunications
server message block (SMB), network file system (NFS), SSH file transfer protocol (SFTP), file transfer
protocol (FTP) for file-based storage and the restful API for object-based storage, etc.
– corresponding protocols or I/O interfaces of a single virtual volume;
– performance acceleration of protocols.
Data manipulation component provides:
– virtual storage pool;
– configuration for virtual storage pool without considering the actual storage location of the data;
– the write buffer or read cache function for customer data;
– enhancement of the read and write response time using high-speed storages for buffer and cache;
NOTE 2 – The high-speed storage includes main memory, non-volatile memory express (NVMe), SSD
and peripheral component interconnect express (PCIe) flash cards.
– data management for snapshots, fast replication and distributed transaction logs.
Data distribution and storing component provides:
– the optimization of writing data to DSF local storage to minimize writing and accessing time;
– data fragmentation to distribute and store in DSF local storage;
NOTE 3 – Data fragmentation is a method to distribute and store customer data in other storages.
– the encryption/decryption and compression/decompression of data fragments;
NOTE 4 – The encryption and compression are taken into account by customer's demands.
DSF local storage management component provides:
– connections to DSF local storage;
– storage tiering of data according to the storage performance, time of data usage and data access
frequency;
NOTE 5 – Storage tiering is the action to distribute and collocate data across multiple storage tiers
in a hierarchical manner.
NOTE 6 – This component automatically moves data between the various storage tiers according to
the characteristics of the data.
NOTE 7 – When data is initially created, it is stored in high-speed storage. When data access
frequency is low, it is moved to a lower-speed storage.
Provisioning and policy management component provides:
– the configurations and controls of logical components;
– the policy management of data storage and data manipulation;
NOTE 8 – Policy for data storage includes back-up, snapshot, scaling, recovery, data caching, thin-
provisioning, tiering, storage type (file, block, object), etc.
NOTE 9 – Policy of data manipulation includes sharing support, read/write, replication, data
migration, fragmentation, encryption, compression, de-duplication, etc.
– the provision of single virtual volume in a virtual storage pool.
Two kinds of metadata are shown in Figure 6-3 as follows:
– Data operation metadata is a description required to perform data operations. It includes the
attributes of virtual storage pool and single virtual volume. It includes a transaction log and DSF data
attributes of read/write caching, snapshot, replication, fragmentation, etc.
– Storage operation metadata is a description required to perform storage operations. It includes the
location of DSF local storage, interface, API for customer data operation, read/write speed, storage
capacity, etc.
258 Network and infrastructure