According to the author, what the value or position is of NoSQL in future data managment?

+ Automatic Zoom BUYER'S GUIDE TO DATA MANAGEMENT I PART 1 OF 3 DATA MANAGEMENT: WHAT DOES THE FUTURE HOLD? T Bernt Ostergaard explores how firms are replacing transactional database management systems with new database architectures ::: 11 Tansactional database management systems OBMS such as IBM DB2 or Microsoft SQL Server have a long and glorious history reaching back to the rule of the mainframe computing paradigm with one server and many users (be they humans or other computing processes). Transactional systems handle the orderly and secure updating and changes to a relational database on a system, when several concurrent transactions are trying to access the same data item, or an outage cuts into an updating process. Queries, updates and data definitions in the DeMS typically use structured query Language (SOU between the user and the DBMS Advantages of the classic transactional DAMS are data inde pendence, efficient data access, data integrity and security, data administration concurrent access, crash recovery OBVIOUS DOWNSIDES However, with the advent of cloud.computing bybrid staan. los and virtualisation, the downsides of transactional DRMS are becoming very obvious Expensive complicated to set up and maintain the relation ship between the data points is essentially constructed at the time of query, and this can be expensive in resource terms. Longer laterky when receiving data from a number of storage locations Software is general purpose, not suited for special purpose tasks (for example, text stream processing) Not good for applications that need real-time processing ENDIMATGETTY HOME 2 of 6 + Automatic Zoom V BUYERS GUIDE Home We can dractors and som them to share Digth Here spending DMLA hts Editor.com NOSQL A data stream management system (SMS) is To manage big data and data volumes that vary Hemaa a set of program comprising input processors, significantly nonlinear databases using Noso, luf continuous query (C) engine and a low-latency (Not Only SQL have been developed by major y grace BRDO cache/buffer conted to backend data stor consumer internet players, such as Google Das bedat age facilities. The DSMS manages continuous data Amazon, Microsoft and Yahoo, that stress scal stre comptes Ranctions on data stream ability NoSQL Des such as Mons. Risk and provides other functionalities as a DUMS Coach and Cassandra rely on specialised DSMS engines such as IBM InfoSplere Stream, frameworks to store data which can be accessed by special SAP Event Stram Processo SAS Event Stream Processing and query APs. This lowers the requirement for consistency to the open source Pipelines are useful for applications that require achieve better availability and partitioning for better scalability processing data streams and require a time or near meal time response with quality of service (0) GRAPH DATABASES Sa instead of querying data assol dors, data is presented to Graphi databases such as SAPH. Nend ArnoB take Queries The produces its own data strum for backend the non-relational data astep further and create a graph of storage, which an SQL for Nosot for unstructured data process relationships. When a query is required to be run against the can then query. The benefits of a coegine are data, the results can be putted out far more efficiently this Accessibility: Live data can be used while still in motion before possible in a relational or basic non relational database, being stored Completeness: Historical data cane streamed and integrated DSMS FOR HIGH VOLUMES with live cata for more cont Users often want to monitor very large data streams, but not High throughput: High velocity, Nigh volume data can be pro actually get notified until specific conditions occur. This requires Cessed with minimal latency incoming data to be treated as a continuous infinite stream of data integrating live and historkal sources, rather than as data COMPLEX EVENT PROCESSING (CEP) in static tables or fides With the arrival of hybrid doud storage, we data and unstruc Typical use cases include network management system, tel tured data in data takes the need to correlate and analyse Infor cos call detall cords, transaction records in francustom mation from multiple and sometimes ad hoc) data sources hors and healthcare monitoring systems increased dramatically. Similarly, much smaller autonomous tuustoot Parcours tuto end wo De Zoom BUYER'S GUIDE Home Why company directors to focus on the small preto was coming Digital amb we driving out in Software spending OVLA des towards a desture Ed units, such as the ones being developed for driverless cars, need to correlate and act on streams of car performance data, real require better predictive capabilities. Software with predictive capabilities are very much ongoing continuous development time sensor data alongside information from the surroundings, processes and rely on algorithms developed by data scientists and data downloaded from external sources, such as traffic and The user community must constantly assess the validity and weather information transparency of the artificial intelligence CEP combines data from multiple CAD process to understand and trust) sources to infer events of patterns that THE GOAL OF COMPLEX EVENT these systems taken together, indicate specific actions. Most of the heavy database suppliers have PROCESSING IS TO DENTIEY ARTIFICIAL INTELLIGENCE products in this category, such as Tibco OPPORTUNITIES OR THEEATS Artificial intelligence encompasses Streambase, Oracle Event Processing SAP wide range of technologies focused on ESP. Microsoft Streaminsight, GigaSpaces ANY RESPOND TO THE AS improving operational efficiency and XAP and Red Hat's BRMS. providing new insights from existing The goal of CEP lo entily opportu QUICKLAS POSSIBLE datat. Al look automate a wide range of nities or threats, and respond to them as processes that may involve reasoning. quickly as possible. These events may happen internally across knowledge planning learning and natural language processing. an artisation, such as sales leads, stock leveb, security Atootliks can be used tobatermachine learning ters, orders or customer service calls, or they may originate from Al applications that are fast and scalable. Another Altool.pl unstructured news items, text messages social media posts, stock is used by developers foron esperiments and make efter pod marketfeeds, trallic and weather reports. The CEP event may also ucts with less trial and error on the marketing front companies be defined as a "change of state," when a measurement exceedsa might employ Conversica, an Artool to priorities customer leads predefined threshold of time, temperature or other we will automated email converses to quality To integrate Al tools with existing data sources, most Als are DEVELOPING PREDICTIVE CAPABILITIES being built primarily by ste igration socialists from the CEPs are limited to historical data and real time information usual suspects such as Accenture down to specialists including Greymruths in the UK and Colent in Germany flows. This does not allow systems to learn from past mistakes Most AF process of data to brand and reprogram themselves to improve their performance. Some fine-tune a system, and remains very much a development processes, such as risk assessments and portfolio Management, Buyer'sgade to data Pared approaches key toge success stalaters won Deme 4 of 6 + Automatic Zoom BUYERS GUIDE we Why company directors now foc on the area SONG Dugtambo are who sowe spend OVLA de las project, primarily driven by industry verticals such as finance with huge amounts of data, millisecond response needs very high a data breach, optical character recognition (OCR), Ieaming to rank, and computer vision transaction volumes and high value and pharmaceuticals, where Machine learning in commercial products.comes from the likes drug depoint efforts require modelling and tralling of wist amounts of data. of Microsoft where Are Lyon CRM service enables users to identify potter over time from MACHINE LEARNING that crop needing time to resto CURRENT, MACHINE LEARNING tion and improving performance At the lowest Al level, process Circo recently announced its automation technoloxy replaces IS MAINLY USED IN COMPLEX Encrypted Italic Analysis (LTA) to find manual handling of repetitive and malware in encrypted traffic Besides high volume tasks. The next level PROBLEMS THAT HAVE PREVIOUSLY the initial datapake in the connection is reached when the Al acquires BEEN HANDLED BY HUMANS EIA looks at the Copacker self-learning capabilities machine lengths and times and the byte distri la) where programs learn to buttons packet payuds within build models by using observations example data) combined allow. The detection process improves over time by expanding with experience data. The resulting models can be predictive or its machine warning models without hoging omsor sew descriptive and continue to evolve by obtaining mowe knowledge trail the last CIA offer uses Nellow data from Chicos about the data at hand Catalyst 2000 switches and 4000 wres Integrated Services Currently, machine learning is mainly used in complex problems Routes integrated with Chea Strath Watch curity analytics that have previously been handled by human, but with no repli cable explanation as to exactly how they solve them and the pro DEEP LEARNING cess of solving the problems is very time and cost-consuming Decoring a specific method of machine Inaring med As opposed to Al, which is based on providing the program with a data intensive, rachine learning processes it relies on GPU rules and user experience machine les employed.com acceleration with for training and interior, and so requires puting tasks where explicit algorithms with good performance are Light integration of hardware and software components in the US Nvidia with its DC line and Voit GPU architecture not available, and where the programming can only provide rules delivers GPU acceleration to detaires desktops Laptop, and forres Example applications include emaileg detec tion of network intruders or malicious Insiders working towards the world's fastest supercomputers. For cloud applications Nvidia kuwee data Pred och key to Based www demand LU? 5 of 6 + Automatic Zoom V BUYER'S GUIDE Home News Why company tractors need to com the pro wang 100 billion neurons and 100 trillion synapses - sa cognitive and lics is still in its infancy. Commercially, the financial sector is at the forefront of cogni. tive analytics adoption and financial services company Opus predicts that in 2017, finance firms in the investment sector will spend $15bn on robotic process automation machine learning deep learning and cognitive analytics, with that sum increasing by 75% to $2.8bn in 2021 Digta ambons uredning growth sowe spending DVLA estos dutta us Editor's NCG deep learning is available on services from Amazon Google IBM and Microsoft. In lapan, Fujitsu announced a deep learning system for Riken, a privately held Japanese research foundation, which in terms of operations will be one of the largest scale supercomputers in Lapan to accelerate research and development in Al technology On the software side, the Google Brain Team has contributed significantly to conduct machine learning and deep neural net- works research with the open source law softw. The flexible architecture runs on one or more CPUsor GPUs in a desk top server or mobile device with a single APL The system indet pins a range of industry-specific deep learning offerings. In the legal field, Intraspexiones TensorFlow as the core of any warning system to investigate and prevenit potentialition COGNITIVE ANALYTICS con analytics relies on processing and store combina tions to mimic the human brain in making deductions from vast amounts of data. This brings us back to the world of the main frame and the supercomputer Recently, the US Air Force and IBM announced a plan to build the world's fast capitis alytics wercomputer to enable analysis of multiple big data feeds (audio, video or text). It will run on an array of North Neurosynaptic processor Each come is part of a distributed network and operates in parallelona event-driven bas not clock-based like conventional CPUs. The process power to 64 min omd6 billion synapses. Estimates of the average human brain indicate hu tata Predapproach ness CAVEATS IN DATA MANAGEMENT Given the breadth of data management is overall maturity is a slow process Driving forces include different technologies platforms and capabilities that influence how this field is prac tised Continued use of SQL and NoSQL for structured data) to access new data storage and data structures is essential to ensuring easy access for developers adoption affordability, and integration with existing enterprise infrastructure. The Nosol scalability is crucials is the ability to enhance corporate data management strategies Likewis, open source developments in Bator big data and Soak for processing will enable companies to expand their data mat tunctionality, while recomb The focus on then be on business issues rather than technology With repositories of structured data especially in datalaks the need for consistent frameworks and data govem ance creases Companies and to develop business males and losses and clearly klentily ace responsability to before making seinic changes to their data management Red w dari orang Denne