Section 214Abstract ± Brooks claims that the very nature ofVofWZaUe iV Whe UeaVon a VilYeU bXlleW haVn¶W beenproduced and though accidental difficulties can beaddressed, the essential difficulties of complexity,conformity, changeability and invisibility (CCCI)means there may never be a complete solution. PriorWo BUookV¶ SXblicaWion, Whe eVVenWial difficXlWieV mahave seemed to be inevitable, however, since thestatement 32 years ago, advances have been made andthis paper aims to find out whether NoSQL is a silverbullet. The discussion in the paper will investigate ifNoSQL can be considered a silver bullet and tacklesthe problems of CCCI.1. INTRODUCTIONoSQL is a non-relational database, designed toaddress performance and scalabilityrequirements of big data and real-time webapplications that traditional relational databasescannot sustain (Ganesh-Chandra, 2015). Incomparisons between MySQL and NoSQL, resultsfrom (Abdullah and Zhuge, 2015) shows thedifference in processing time between the twodepending on the number of objects provided,whilst 1000 objects show a 0.012-time difference,this changes to 16.220 seconds when 1000000objects are implemented, MySQL 18.620, NoSQL2.400. This vast difference is an accomplishment ofNoSQL and how it has changed the approach to bigdata. Although can it be considered a solution tothe issues of CCCI?When discussing the complexity of software,it is important to acknowledge that scaling-up of asoftware entity is not represented by the sameelements just in larger size but rather an increase ofthe number of different elements, resulting in theseelements interacting with unpredictable behaviour.This can also be applied to the complexity ofNoSQL which this paper will investigate further.WhilVW cRQfRUPiW iVQ¶W XQiTXe WR VRfWZaUe,the standards that is expected differ greatly. Inmathematics, new algorithms might be designed tooutperform previous models, but the fundamentalsstay the same. In software, there needs to beconformity depending on the interface,specifications and laws which can change over aSeUiRd Rf WiPe. If Whe VRfWZaUe dReVQ¶W PeeW WheVerequirements, it could be considered asunprofessional quality or simply a failure. Whenimplementing a NoSQL database, it does not needto conform to a structure like a relational databaseneeds to, however, there would need to beconformity when analysing the data for the client tomake use of it.There is a variety of reasons software mightneed to change in the current environment. Demandfor innovation from consumers, change in legalityor software unable to keep up with technologyadvances. Regardless of the reason, there is anever-increasing number of consumers, Facebookhas an estimated 100 billion messages shared perday (About Facebook, 2019). The modern systemneeds to handle high amounts of data in new waysthat traditional methods cannot withstand, and withdata warehouses being indispensable (Mehmoodand Anees, 2019) software engineers and thesystem need to keep up with the change.Unlike most professions, software systemsare invisible, code is not a physical object, nor canthe designs be constructed in the same manner asavailable to an architect. Though tools like ERDand schema make a representation that softwareeQgiQeeUV caQ agUee RQ, bXW Whe dRQ¶W cRPSUeheQdthe final product. In systems that are large orabstract the diagrams are more complex but stillneed to conform with the current standard andadapt to change when required whilst still beinginvisible.2. DISCUSSION2.1. ComplexityThere has been research carried out to tackle theproblem of complexity in NoSQL. The workconducted by Corbellini (2017) to determine whichmethod of NoSQL is most viable when usingAQ iQYeVWigaWiRQ iQWR NRSQL aQd cRQVideUiQg WheiQheUeQW difficXlWieV iQ SRfWZaUe EQgiQeeUiQg.Sam Town, 0009630977, Software Engineering, [email protected]NKey Words ʹ NoSQL, Complexity, Conformity, Changeability,Invisibility, Silver Bullet, Essential DifficultiesSection 215bigdata, suggests that as NoSQL databases are abroad spectrum (Corbellini et al, 2017), therewould be a solution to the majority of problems incurrent database systems although, Corbellini(2017) also proposes that not every applicationrequires a NoSQL database due to the complexityinvolved (Corbellini et al, 2017). Mior (2017)continues the discussion of complexity regardingschema designs for NoSQL applications (Mior etal, 2017).Although it is important to choose a goodschema, there is not an established methodology ortools to guide this process (Mior et al, 2017)leading to the design of NoSQL databasesremaining a complex issue in comparison torelational databases (Mior et al, 2017). Thissuggests the complexity of designing andimplementing a NoSQL database is still acontinuous problem within the industry.In contrast to this Ramzan et (2019), impliesthat the rise of bigdata has led to NoSQL being themost viable database solution and the complexity isnot due to NoSQL but rather a result of relationaldatabases needing to shift to the flexible database(Ramzan et al, 2019).However, the study by Ramzan (2019) investigatedthe migration from traditional relational databasesto NoSQL and did not expand on the complexityfor programmers to understanding or create aNoSQL database.2.2. ConformityThere is a need for programmers to conform to astandard, this leads to a collective understandingand reducing the complexity of software design,this should also apply to NoSQL databaseshowever, though Miro (2017) agrees of theimportance of a good schema design, they argueµWheUe aUe QR WRRlV RU eVWabliVhed PeWhRdRlRg WRgXide aQd VXSSRUW Whe SURceVV¶ (Miro et al, 2017),leaving the design to previous experience based onpractice rather than theoretical understanding.A study by Lucchese and Henrique (2017)was done in order to understand the suitability ofNoSQL storage systems dependent on theapplication requirements by using a metric system.µBig DaWa iV a UaWheU QeZ cRQceSW aQd iWV ecRVVWePof tools and applications has been subject tocRQVWaQW WUaQVfRUPaWiRQ¶ LXccheVe aQd HeQUiTXe(2017). This suggests that as Big Data and NoSQLdevelops, the tools available will also progress andlead to conformity by software engineers.Although, even with more advanced tools, thiswould not mean NoSQL has dealt with the issue ofconformity completely as expressed by Schreiner(2019), there is not a standard query language forNoSQL leaving a steep learning curve fordevelopers.2.3. ChangeabilityNoSQL is the result of a need for change thecurrently used methods, applications have seen theamount of data exponentially grow in recent years(Abdullah and Zhuge, 2015). The success ofNoSQL databases has also led to hybrid methodsbeing created as programmer need to move beyondthe original domain.Work by Schreiner (2019) also investigatesthis, by implementing relational schemata and SQLinstructions to equivalent schemata and usingaccess methods of any key-oriented NoSQLdatabase (Schreiner et al 2019) would reduce theoverall processing time. The results from this studyshowed a reduction in processing time and a 95%confidence interval (Schreiner et al 2019).Research from Wu (2017) also investigates thenotion of a hybrid NoSQL-SQL database inrelation to real-time geospatial data, showingimprovements in update and query speed whilstreducing latency when compared to MongoDB andMySQL (Wu et al, 2017). This suggests thatNoSQL is not stagnated and can change from itsoriginal design, however, though this could beconsidered successfully in Brooks terms, there isstill the issue of the complexity of hybrid versionsand not conforming to an agreed upon standard.2.4. InvisibilityNoSQL is non-relational, not constraint in the sameways as traditional SQL, allowing the developermore autonomy when designing the database.However, Zhang (2019) argues that since there arefew rules and restrictions, the design of storageschema for NoSQL can be difficult, as there couldbe a lack of understanding or experience by thedeveloper.Sellami and Defude (2018), also agreed onthis point, claiming that as programmers need to befamiliar with data models of different applicationprogramming interfaces (APIs), complex queriescannot be achieve in a declarative way and needsextra implementation effort (Sellami and Defude,2018).3. ConclusionNoSQL helps software engineers in storing largeramounts of real-time data more dynamically incontrast to traditional methods, although, as thedata does not need to be defined, it can be arguedthe structure of NoSQL is of less-quality in thisregard. Ganesh (2015) expresses the downsides toNoSQL are its lack of a universal language, eachNoSQL product is unique and relational databaseshave matured over the past 40 years allowing for anecosystem of applications and tools available.In terms of CCCI, NoSQL does not reduceWhe cRPSle[iW aV ³PaQ cRPSaQieV aQd RSeQSection 216source communities promoting its ownimplementation, making the use of NoSQLtechnology in a production environment a toughdeciViRQ´ CRUbelliQi eW al (2017). ThiV VhRZV WhaWthe variety of NoSQL databases also contributes toits complexity.Conformity continues to remain an issue, thelack of a universal language and a need forestablished methodologies means both conformityand complexity cannot be fully solved.In terms of transitioning a relationaldatabase to a NoSQL database, further researchneeds to be done to reduce the complexity andmaking change more intertwined with the culturalmatrix of software engineers. However, the use ofNoSQL in hybrid applications shows the potentialof challenging the issues of changeability.Due to the very nature of software, invisibilitywill remain an issue as it is not something physical.With more conformity, leads to more changeability,then leads to more complexity and increasesinvisibility.4. EvaluationNoSQL shows the progression of the current needof databases and though it makes improvements inhandling bigdata, it does not deal with the fourissues of software expressed by Brooks. Forexample, the metric system on NoSQL suitabilityby Lucchese and Henrique (2017) takes a step indealing with conformity, nonetheless without astandard query language, NoSQL will alwaysremain complex (Schreiner et al 2019). There arebenefits of using NoSQL but as negativities remainin understanding and implementing the software,the concept of a silver bullet cannot be applied toNoSQL as it would need to eradicate all negativeissues regarding to the CCCI.5. Future WorkFuture work on this subject could include furtherinvestigation on creating a standard query languagetool, Mahajan, Blakeney and Zong (2019) havedone research into improving the energy efficiencyof NoSQL databases via using query optimizationshowever, not enough research has been done intothe theoretical side of implementing a universalstandard query language.6. REFERENCES[1] G Abdullah, A. and Zhuge, Q. (2015) FromRelational Databases to NoSQL Databases:Performance Evaluation, Research Journal ofApplied Sciences, Engineering and Technology,11(4), pp. 434-439.[2] About Facebook. (2019). Company Info | AboutFacebook. [online] Available at:https://about.fb.com/company-info/ [Accessed10 Nov. 2019].[3] Brooks (1987). No Silver Bullet Essence andAccidents of Software Engineering. Computer,20(4), pp.10-19.[4] Corbellini, A., Mateos, C., Zunino, A., Godoy,D. and Schiaffino, S., 2017. Persisting big-data:The NoSQL landscape. InformationSystems, 63, pp.1-23.[5] Ganesh Chandra, D. (2015) BASE analysis ofNoSQL database, Future Generation ComputerSystems, 52, pp. 13-21.[6] Lucchese, F., 2018. From P2P to NoSQL: acontinuous metric for classifying large-scalestorage systems. Journal of Parallel andDistributed Computing, 113, pp.227-249.[7] Mahajan, D., Blakeney, C. and Zong, Z. (2019)Improving the energy efficiency of relationaland NoSQL databases via queryoptimizations, Sustainable Computing:Informatics and Systems, 22, pp. 120-133.[8] Mehmood, E. and Anees, T., 2019.Performance Analysis of Not Only SQL SemiStream Join Using MongoDB for Real-TimeData Warehousing. IEEE Access, 7,pp.134215-134225.[9] Mior, M.J., Salem, K., Aboulnaga, A. and Liu,R., 2017. NoSE: Schema design for NoSQLapplications. IEEE Transactions on Knowledgeand Data Engineering, 29(10), pp.2275-2289.[10]Ramzan, S., Bajwa, I.S., Ramzan, B. andAnwar, W., 2019. Intelligent Data Engineeringfor Migration to NoSQL based SecureEnvironments. IEEE Access.[11]Schreiner, G., Duarte, D. and dos SantosMello, R. (2019). Bringing SQL databases tokey-based NoSQL databases: a canonicalapproach. Computing.[12]Sellami, R. and Defude, B. (2018) ComplexQueries Optimization and Evaluation overRelational and NoSQL Data Stores in CloudEnvironments, IEEE Transactions on BigData, 4(2), pp. 217-230.Section 217[13]Wu, C., Zhu, Q., Zhang, Y., Du, Z., Ye, X.,Qin, H. and Zhou, Y. (2017) A NoSQL±SQLHybrid Organization and ManagementApproach for Real-Time Geospatial Data: ACase Study of Public Security VideoSurveillance, ISPRS International Journal ofGeo-Information, 6(1), p. 21.[14]Zhang, D., Wang, Y., Liu, Z. and Dai, S.(2019) Improving NoSQL Storage SchemaBased on Z-Curve for Spatial VectorData, IEEE Access, 7, pp. 78817-78829.
