IBM Spectrum Storage Suite

In IBM automation of data storage began since 1988, with the product Workstation Data Save Facility (WDSF) which, as it appears from the name, intended for workstations. In 1996 IBM purchased Tivoli then with the product IBM Tivoli Storage Manager it was succeeded to lift automation of storage to system level. The IBM Tivoli Storage Productivity Center system released later became a prototype of the product IBM Spectrum Storage Suite for the first time provided in 2015.

Features of approach to management of DWH

The complex of products for management of DWH offered IBM as of 2017 is not accidentally called by IBM Spectrum Storage Suite, it really connects in a uniform range all other products of corporation intended for data storage. In its basis software-defined approach.

The fact that at the end of 2016 the IDC analytical company recognized the third year in a row IBM the leader in the field of software-defined platforms of data storage (Software Defined Storage Platforms – SDS-P) demonstrates to perfection of set of IBM Spectrum Storage Suite technologies.

In the report of IBM Software-DefinedStorage Guide software-defined DWH is considered as one of three software-defined components of infrastructure along with computing and network resources (Rice 1).

Fig. 1. Three software-defined components of infrastructure

Storage Suite differs in complete representation of resources of corporate DWH and, as a result, an opportunity to place data where it is the most reasonable. IBM Spectrum Storage Suite allows to work more effectively with data that increases flexibility and business performance. Are among the main advantages of Spectrum Storage:

• Simplification and integration of management of storage and data protection for traditional and new applications
• Elastic scalability and high performance during the work with analytics, Big Data, social and mobile technologies
• Elimination of separation for delivery of data "without borders" due to the built-in support of hybrid cloud environments
• Optimization of economic indicators of work with data due to intellectual separation into levels of flash-storages, tape drives and cloud environments
• Use of open architectures with support of different industry standards, including OpenStack and Hadoop

By means of IBM Spectrum Storage it is possible to manage on a centralized basis more than 300 different data storage devices. Application of IBM Spectrum Storage allows to lower storage costs of data, automatically moving information on the most economic device.

IBM Spectrum Storage Suite implements separation, traditional for SDS, into two Control Plane and Data Plane levels where Control Plane focuses in itself all control functions by the devices located at the level of Data Plane. Data Plane supports direct interaction with drives among which can be as released by IBM, and other producers.

At the level of Control Plane control of the virtualized resources is exercised, the functions of the high level necessary for support of loadings and also are provided to the user for ensuring optimization, scaling and fast deployment of infrastructure of storage. This level is generally formed of the products entering earlier into the Tivoli family. Now they are called Spectrum Control, Copy Services Manager, Spectrum Protect, Spectrum Protect Snapshot and Spectrum Copy Data Management.

At the level of Data Plane such control functions of DWH as virtualization, support, RAID, movement on levels of storage (tiering), enciphering, deduplication are performed. Here the interface with devices, including block I/O (iSCSI, etc.), file (compatible to POSIX) and Hadoop Distributed File System (HDFS) is supported. This level is formed by the products Spectrum Virtualize, Spectrum Accelerate, Spectrum Scale and Spectrum Archive.

Rice 2. Separation of management into levels

Problem of a new solution is change of the organizational and economic principles of system management of data storage and adding in it the intellectual software, speak in IBM:

New approach which could help customers to cope with the increasing cost and complexity of data storage because of their increased volume is necessary. In the modern world traditional methods are already inefficient, especially when the value of each element constantly changes. IBM radically changes idea of storage systems thanks to Spectrum Storage technology which will allow customers to distribute more effectively the equipment costs for the purpose of extraction of commercial value from data

Range of hardware solutions

Four families disk IBM XIV, IBM DS8880, IBM Storwize and FlashSystem Family in combination with tape libraries TS4500 and TS3500 form the complementary complex of technical means integrated in software-defined infrastructure. On each of families informative works from the IBM Red book series in the section of Storage are written.

IBM XIV Storage System

IBM XIV Storage System conducts the family tree from the drive Nextra purchased together with company XIV in 2005. In Nextra the classical problem of an engineering compromise was solved - as from mass, cheap and not too reliable components to collect a reliable and productive system. The originality of the idea consisted in use of the disks SATA.

The disk SATA can execute 100 transactions per second that it is twice less than at disks of SAS, and the delay is twice more. But at reasonable replacement of 100 disks of SAS by 200 disks SATA we can receive the same joint speed of exchange. Construction which will be able to implement necessary conditions is important. Besides, the disks SATA, at least, are three times cheaper, six times more on volume and consume energy twice less. As a result, having spent three times less than means, it is possible to receive six times bigger storage which besides is twice more economical regarding an electric power expense.

IBM XIV Storage System is constructed on cluster to the scheme with orientation to horizontal approach to scaling (scale out) — it is possible to increase the capacity of disks, the size a cache memory, bandwidth between disks and a cache and between a cache and hosts, the processor power of interface units and modules of data, processor power a cache memory.

The data placed in XIV "are cut" counting on all available spindles that provides high performance at low price. If to transfer to terms of three-level model, one may say, that the quality of the Tier 1 level is available at the price of Tier 2. All disks become virtualized in a uniform pool in which a part of disk space is left for replacement of faulty disks, and remained shares equally between the main and reserve copies of data.

Formatted capacity of IBM XIV Storage System is up to 2 PB on a system and 288 PB when using 144 systems (using hyper scalable management tools).

DS8880

DS8880 family does not need special representation. In 1998 IBM released the type of the equipment IBM Enterprise Storage Server (ESS) new for those times performing function of the corporate server of data, it is known as Shark. In 2004 upgraded Shark was renamed into DS8000. So there was a family constructed on Power 5 processors. These drives had to compete with Hitachi TagmaStore and EMC Symmetrix. As of 2017 in several options DS8880 - the system with scaling of scale up providing support of different platforms including IBM z Systems mainframes is issued:

• DS8884 is a business class
• DS8886 is a corporate class
• DS8888 is an analytical class

All three are constructed on the IBM Power 8 processor, DS8884 and DS8886 options can be completed with the hard drives HDD, and in combination with SSD to become hybrid, or to be completed only SSD, i.e. to turn into an array of AFA (All Flash Array). DS8888 model rehearsed on analytics is completed only with the disks SSD.

IBM Storwize V7000

In 2010 Storwize owning technology of a compression of data in real time was purchased. It allows to squeeze all types of data which are at the different levels of storage using an algorithm Random Access Compression Engine (RACE), without having an impact on operation of applications. There are two major versions of the solution - IBM Storwize V7000 and IBM Storwize V7000U and also several designs (V5000, V3500 and V3700) differing on a complete set in disks. Index U indicates the version combining block and file data access. A system combines properties of the SAN (iSCSI/FC) and NAS systems. Versions of Storwize V7000F and Storwize V5030F are completed only a flash disks of SFF (small form factor).

IBM Storwize V7000 differs in simplicity in operation, performance and also quite compact sizes. This system is quite available to the majority of averages by the size of the companies. In spite of the fact that IBM Storwize V7000 is the system of the middle class, it includes the latest advanced technologies from IBM which were used only in DWH of a corporate class earlier. IBM Storwize V7000 supports the 2nd terabyte disk 3.5’’ and also disks – 300, 450, 600 GB 2.5’’ and 300 GB the SSD solid state drive. The maximum number of hard drives in a system – 240, volume a cache memory – 8 or 16 GB.

FlashSystem Family

After acquisition in 2013 of Texas Memory Systems, the company specializing in DRAM and SLC technologies for IBM an opportunity to release two products of the FlashSystem family - simpler specialized system (appliances) 900 Series and V9000 complemented with the SAN controller known as SVC (SAN Volume Controller) opened. The Real-time Compression technology gives almost fivefold gain of formatted capacity of data storage.

IBM FlashSystem V9000 differs in advantages of a software-defined storage system and high-speed performance a flash memory. In it the high performance and short delays are combined. V9000 has rich feature set, including data compression in real time, runtime allocation of data on storage levels, economical selection of resources, service of copying of data and a configuration of high availability.

V9000 increases performance of applications and infrastructures. It replacing with itself a set of racks with hard drives considerably cuts down expenses on the electric power, lease of premises and cooling, increases efficiency of servers, cuts down expenses on purchase of licenses for software and also reduces expenses on power supply and cooling.