Comparing IBM DB2 for z/OS’s High Performance Analytics and Transaction Processing to Oracle Session #6973 Sheryl M. Larsen, IBM Wed, 29/Oct, 03:00 PM - 04:00 PM © 2014 IBM Corporation Please Note • IBM’s statements regarding its plans, directions, and intent are subject to change or withdrawal without notice at IBM’s sole discretion. • Information regarding potential future products is intended to outline our general product direction and it should not be relied on in making a purchasing decision. • The information mentioned regarding potential future products is not a commitment, promise, or legal obligation to deliver any material, code or functionality. Information about potential future products may not be incorporated into any contract. • The development, release, and timing of any future features or functionality described for our products remains at our sole discretion. Performance is based on measurements and projections using standard IBM benchmarks in a controlled environment. The actual throughput or performance that any user will experience will vary depending upon many factors, including considerations such as the amount of multiprogramming in the user’s job stream, the I/O configuration, the storage configuration, and the workload processed. Therefore, no assurance can be given that an individual user will achieve results similar to those stated here. 2 Three Myths #1 Oracle RAC can scale like DB2 #2 Oracle has the same availability as DB2 #3 Oracle Exadata Database Machine can do business analytics like DB2 + IBM DB2 Analytics Accelerator Oracle Real Application Clusters for OLTP Demo Server 1 Server 2 Database Resources (rows) Transactions Oracle RAC for OLTP Server 1 Server 2 Database Resources (rows) Transactions Oracle RAC for OLTP Server 1 Transactions Server 2 Oracle RAC for OLTP Server 1 Server 2 Inefficient utilization of resources! Limited Scalability! Transactions Oracle RAC uses network based sharing Shared disk across nodes Software-based distributed lock and cache management InfiniBand Shared Interconnect Less efficient scalability due to increasing inter-node activity as cluster grows or as workload increases Locks Cache Locks Cache Oracle Oracle Unable to scale bandwidth for Data Warehouse processing Locks Cache Oracle Switch Disks BEFORE / WITHOUT EXADATA IBM Internal Use Only 8 Oracle 12c with Real Application Clusters (RAC) • 3 Oracle 12c Instances, 1 per server • RAC is software that manages the application clusters across the servers High Speed Interconnect Separate Servers - No shared components - Each running an Oracle Instance Fiber Channel Switch - for concurrent access to shared storage Single storage with single copy of the database 9 DB2 for z/OS Data Sharing is hardware based • There are many ways to separate DB2 Databases For example, 1 LPAR on each machine Coupling Facility (dedicated processors and memory) Group Bufferpools Centralized Lock Mechanism Shared Storage DB2 for z/OS delivers a truly scalable shared disk architecture True shared caching Delivers better availability Hot pages stored in centralized cache Faster recovery if one member fails 10 Myth #1 Oracle RAC can scale like DB2 DB2 Parallel Sysplex Clusters vs. Oracle RAC DB2 for z/OS Oracle RAC Centralized Coupling Facility Design Distributed Design CF Locks Cache Network Switch InfiniBand Switch DB2 DB2 DB2 Member Member Member Shared Disk lock and buffer management communicates with central coupling facility Locks Cache Oracle Node Locks Cache Oracle Node Locks Cache Oracle Node Shared Disk distributed locking and buffer management 11 Oracle RAC’s distributed lock management design causes network overhead X X 115 100 100 Master Node Instance Instance Instance Instance A B C D Network Transfer Lock Inform Request Release Done assume Lock Lock Lock Write Holder and Lock Data Inter-node connections: 1 2 3 4 5 6 7. A 1. 2. 3. 4. 6. D B C releases updates requests request acknowledges C haveAdata, local write C read to release copy lock 5. and A lock and locks lock assumes to release and and local transfer lock lock copies data transfers data element D is element and theinforms locktomaster BD for this element In a cluster with 4 nodes, an update operation may need 6 network connections and two in-memory calls (not shown). Example based on Oracle’s US Patent 7,107,319 B2. 12 zEnterprise’s centralized Coupling Facility permits efficient lock and cache management in DB2 z/OS image B with DB2 z/OS image A with DB2 X 100 CF can support over 1 million requests per second 100 115 115 CF Buffer Pools Locks High Speed Hardware Links 1. B registers page to CF and obtains write lock 2. B updates data 3. B commits update B caches update in group buffer pool X 100 z/OS image C with DB2 A and C have data in local buffer pool without locks z/OS image D with DB2 CF invalidates all cached copies without interrupting processors Cache and locks are maintained with no inter-node disturbance! 13 Myth #1 Oracle RAC can scale like DB2 DB2 Parallel Sysplex Clusters Scale far Better! DB2 for z/OS Near Linear Scalability • DB2 for z/OS - Adding nodes is close to perfect linear scaling 12 Perfect Linear Performance Productive Nodes 11 Productive Nodes 10 9 9.98 8 7 6.86 6 5.24 5 4 3.48 3 2 1.78 1 1 2 3 1.69 4 Oracle RAC Poor Scalability 2.44 5 6 7 8 Nodes in Cluster 9 10 11 • Oracle inter-node communications make scaling difficult 12 Oracle RAC characteristics as shown in Dell RAC InfiniBand Study http://www.dell.com/downloads/global/power/ps2q07-20070279-Mahmood.pdf CERN (European Organization for Nuclear Research) http://www.oracleracsig.org/pls/apex/RAC_SIG.download_my_file?p_file=1001900 Insight Technology http://www.insight-tec.com/en/mailmagazine/vol136.html DB2 for z/OS OLTP result (ITG ‘03) 14 What about Availability? Myth #2 Oracle RAC can have the same availability as DB2 zEnterprise EC12 Parallel Sysplex Availability Leads The Pack Downtime Survey (400 participants in 20 countries) Downtime Hours Per Year 3.5 3 System z Parallel Sysplex with 99.999% availability 5 minutes 2.5 2 1.5 1 0.5 0 Microsoft Win2003 on Intel x86 Microsoft Win2008 on Intel x86 Novell HP HP Linux on HP-UX HP-UX Intel x86 on on (unmodified) Intel Itanium HP PA-RISC Sun Solaris on Sun SPARC Mandrake Linux on Intel x86 Novell Linux on Intel x86 (modified) IBM AIX on IBM Power IBM z/OS Sysplex Source: ITIC: ITIC 2009 Global Server Hardware & Server OS Reliability Survey; July 2009; http://itic-corp.com/blog/2009/07/itic-2009-global-server-hardware-server-os-reliability-surveyresults/; Results are measured in minutes per year. *Note: All operating systems included in the survey are not included in this chart. Fifteen operating systems on various processor architectures were included in the survey. The chart will be updated when the full report is available. 16 Myth #2 Oracle RAC can have the same availability as DB2 Availability: Maintenance and Upgrade DB2 for z/OS: Applications keep running during maintenance and upgrades! Maintenance in place even with mixed versions Refresh in place Oracle RAC requires downtime to upgrade the database Source: Fact that RAC required downtime to upgrade is documented in Metalink Note 244241.1 zEnterprise parallel sysplex clusters provide unmatched processing power and availability • Clustering driven by specialty engines (Coupling Facility) • Presents a single system image of a z/OS workload CF CF Single System Sysplex CF CF • Potentially 2.5M MIPS per 32-way cluster* *Equivalent to about 240 of the largest Oracle servers CF CF External Coupling Facility (Can be different class server) • Enables rolling updates CF CF Cross Connected Servers with internal Coupling Facilities • Supports continuous access to business services and data – from anywhere, at anytime • Designed for 99.999% availability 18 Myth #3 Oracle Exadata Database Machine handles business analytics like DB2 + IBM DB2 Analytics Accelerator Optimal Execution of Mixed Workloads Transactional Many concurrent transactions throughput Business Analytics Deep Analytics Complex analytic queries response time Operational Analytics Many concurrent queries of varying complexity throughput Real time answers cannot be derived from extracted data sources 19 Exadata to the Rescue? • Exadata intended to “fix” Oracle storage bottleneck for DW Exadata Storage Servers Networking Database Servers For every Oracle database server there are additional “Exadata” servers to perform the I/O, each with 12 disk drives and new Exadata software • Extreme Performance and Scalability for All Applications • Oracle Exadata X4-8 is especially well suited for high-end OLTP workloads, in-memory or memory-intensive workloads, large-scale database consolidations, including DBaaS, and multi-rack data warehouses. IBM Internal Use Only Exadata is a Hardware Band-Aid for Oracle RAC • Still uses Oracle RAC at the end of the day Oracle is simply throwing lots of H/W at this problem to make it faster • Another layer of servers which all need to be licensed • Oracle needs this band-aid because they do not have an MPP solution MPP is the proven way to properly reduce I/O (joins, aggregates, query push downs, etc.) and interconnect traffic • Adding another layer of servers to scan and filter data still doesn’t solve RAC’s problems Example: “large sorts cause massive performance problems with RAC” • Quote from noted Oracle expert, Don Burleson Internal Use Only What They are Advertising Oracle Exadata Database Machine is best for: • ERP, CRM, data warehousing, HR • Mix OLTP, DW, batch and reporting • Large memory allows many databases to be consolidated 22 Exadata Adds an I/O Layer and Complexity to RAC Requires Oracle 11g R2 or higher Enterprise Server plus Oracle RAC software Limited to predicate and column filtering for table scans and large-table-small table join processing. Reality – Managing Parallel Execution of Concurrent Workloads With Oracle RAC is EXTREMELY COMPLEX! Internal Use Only The Exadata Database Machine summary • A hardware band-aid for Oracle RAC to “fix” I/O performance issues • • • • Is Not Smart Storage – just a bunch more servers Is complex under the covers – 2 layers of server clusters Degraded performance upon server failure Not Sparc and Not Solaris* (runs Oracle Enterprise Linux, which is not widely accepted in the market) • Exadata Smart Scan disrupted when data is actively being updated • Exadata compression enabled only for bulk loading read-only data • Doesn’t include OLAP, data mining, business intelligence • Newest edition has slower discs increasing I/O from 4 – 6ms 24 What Else are They are Advertising? • Oracle Exalytics In-Memory Machine X3-4 Engineered System for Extreme Analytics Oracle Exalytics is the industry’s first engineered system for inmemory analytics that delivers extreme performance for Business Intelligence and Enterprise Performance Management applications. Built using best in class hardware, market- leading business intelligence software and in-memory database technology, Oracle Exalytics is an optimized system that delivers speed of thought analysis with unmatched intelligence, simplicity and manageability. Requiring no application redesign Bundle of Oracle Business Intelligence Foundation (Oracle OBI EE, Oracle publisher, and Essbase) together with Oracle’s Times Ten in-memory database Exalytics Sun Fire x4470 M2 server. 25 Oracle Summary • RAC lives inside Exadata • Data often arrives from DB2 Once the data is extracted it is old and stale unless kept in sync with replication Exadata exports data to Exalytics Exalytics Hosts a whole database, data mart, result cache or cube in 1TB max memory 26 They adopt an extremely expensive ETL strategy to support analytics A large European bank: Analytical Data Analytical Data Analytical Data Analytical Data Operational Data Analytical Data 120 database images created from bulk data transfers 1,000 applications on 750 cores with 14,000 software titles ETL consuming 28% of total distributed cores and 16% of total MIPS A large Asian bank: One mainframe devoted exclusively to bulk data transfers ETL consuming 8% of total distributed core and 18% of total MIPS With this strategy, IT costs grow faster than business growth Source: IBM Eagle Studies 27 What about Security? System z – Protect Critical Data End To End Privacy Rights Clearinghouse noted 621 security breaches during 2013, involving 258 million sensitive records … System z security breaches: Source: http://www.privacyrights.org/data-breach/new http://www.redspin.com/docs/Redspin-2013-Breach-Report-Protected-Health-Information-PHI.pdf http://www.forbes.com/sites/maggiemcgrath/2014/01/10/target-data-breach-spilled-info-on-as-many-as-70-million-customers/ 29 0 Oracle security patches far outnumber those for DB2 for z/OS 100 90 Cumulative Database Security Fixes Since 1Q10 Number of Fixes 80 Oracle: 14 security patches during the past year: April 2014 - 104 total, 2 for the database January 2014 – 144 total, 5 for the database October 2013 Oracle DB - 127 total, 4 for the DB2 database on z/OS July 2013 - 89 total, 3 for the database 70 60 Oracle DB DB2 for z/OS 50 40 30 20 10 0 2010 2011 2012 2013 DB2 – only 40 security patches over past 29 years Source: http://www.oracle.com/technetwork/topics/security 30 Back to Myth #3 Oracle Exadata Database Machine handles business analytics like DB2 + IBM DB2 Analytics Accelerator IBM has the only near-Real Time Analytics! Business Analytics Operational Analytics Deep Analytics Operational Data Analytics Data IBM zEnterprise Analytics System 9700 • A workload-optimized, blade-based appliance that runs queries in seconds versus hours Storage integrated into the hardware rack Eliminate table indexing and query tuning • Deeply integrated with DB2 for z/OS, and transparent to applications Pre-load data from DB2 for z/OS into Accelerator at over 400GB/hr zEC12 IBM DB2 Analytics Accelerator Maintain a single copy of table or partition in Accelerator and update incrementally (High Performance Storage Saver) System z workload management implemented across Accelerator 32 Oracle RAC for OLTP Server 2 Server 1 ` Inefficient utilization of resources! Limited Scalability! Transactions The Analyst Community Has Taken Notice! • “By eliminating analytic latency and data synchronization issues, hybrid transaction/analytical processing will enable IT leaders to simplify their information management infrastructure” “This architecture will drive the most innovation in real-time analytics over the next 10 years via greater situation awareness and improved business agility” Real Time Analytics that minimizes or eliminates analytics latency or synchronization issues by eliminating the divide between operational and analytical systems. Hybrid Transaction and Analytics Processing (HTAP) Gartner Research Note G00259033: Gartner 01-2014 Hybrid Transaction Analytical Processing Will Foster Opportunities The best-fit solution – Move analytics closer to the data Extract Transform Load (ETL) Data Store Analysis Tools InfoSphere Information Server DB2 Cognos BI z/VM z/OS z/VM or z/OS IBM zEnterprise Analytics System 9700 – A comprehensive packaged solution including hardware, OS, and business analytics software SPSS z/VM Full function operational business intelligence AND business analytics on the same platform FastStart Service Pack enables quickest time to value with the least amount of impact Data Integration Pack provides data movement and transformation, data discovery and real-time delivery Data Analytics Pack includes QMF, Cognos and SPSS Predictive Analytics zEnterprise IBM DB2 Analytics Accelerator 35 That’s what IBM did! Blue Insight delivers incremental value across IBM Building a “culture of performance” by Progressing the Possible to the Proven “World's Largest Cognos Deployment” Solution Components • • • • • • • IBM® Cognos® Business Intelligence IBM SPSS® Statistics IBM SPSS Modeler IBM InfoSphere® Warehouse IBM zEnterprise™ 196 IBM DB2 for z/OS IBM DB2 Analytics Accelerator Technical Value 100+ stand alone deployments into ONE analytics environment Supports 450,000+ global users Consolidates value from 660+ data warehouses 500+ analytic applications across all business processes Business Value Produces 5+ Million $ in yearly efficiency savings Conservatively added 300+ Million in business value this year IBM Software Computer Services IBM Blue Insight – Turning big data into actionable business insight – fast and at low cost 400 percent faster responses to complex queries 100 hours to 6 minutes acceleration for one key analytics report Game-changing insight available almost in real time Solution components The transformation: Previously, analyzing huge volumes of data from hundreds of systems required the creation of multiple inflexible data marts. Today, IBM can deliver near-real time analytic insight without going through manual data extraction and preparation stages, enabling users to respond to emerging business challenges faster and make more timely, fact-based strategic decisions. Software • IBM® DB2® Analytics Accelerator for z/OS® “DB2 Analytics Accelerator has enabled us to change the • IBM Cognos® Enterprise • IBM DB2 for Linux, UNIX and Windows way we do business.” —Larry Yarter, Chief Architect, IBM Blue Insight Business Analytics • IBM DB2 for z/OS Competency Center • IBM PureData™ for Analytics • IBM SPSS® Statistics • IBM SPSS Modeler Hardware • IBM zEnterprise® 196 ZSP03876-USEN-02 System z: A Hybrid Transaction and Analytics Platform HTAP A single workload-optimized system for accelerating decisions to the speed of business Integrate DB2 11 and Big Insights provide this extended capability OLTP Transactions Operational analytics High concurrency DB2 Native Processing Advanced analytics* Standard reports OLAP Complex queries Historical queries IBM DB2 Analytics Accelerator Real time data ingestion Everything is online – analytics in the right place! Data is our clients’ most important resource. The more effectively and efficiently they use that data, the greater their competitive advantage. Analytics models and software. These are the tools that deliver actionable insights from data. Predictive View Data warehouses, marts, etc. These data sources support reporting and predictive model creation. Historical View The operational systems that house the book of record. These data sources are critical to dayto-day business processes. Real-Time View 39 A D V A N T A G E … what happens if trends continue? … what happens if <fill in the blank>? … what is likely to happen next? … how to achieve the best outcome? … what happened? … how many, how often, where? … what the problem is, exactly? … what actions are needed? Business-critical analytics solutions depend on where the source data resides DB2: Compare What is happening now or in the future analytics? Incremental updates keep Data in sync B u s i n e s s V a l u e Oracle: What happened at the time of extract analytics? Exalytics A static data environment Business Value 40 Leading The World For Hybrid Transaction & Analytic Processing, HTAP! IBM zEnterprise #1 Efficiency at scale #2 Trusted resiliency #3 Optimized for OLTP and near real-time analytics zEC12 IBM DB2 Analytics Accelerato r 41 Acknowledgements and Disclaimers Availability. References in this presentation to IBM products, programs, or services do not imply that they will be available in all countries in which IBM operates. The workshops, sessions and materials have been prepared by IBM or the session speakers and reflect their own views. They are provided for informational purposes only, and are neither intended to, nor shall have the effect of being, legal or other guidance or advice to any participant. While efforts were made to verify the completeness and accuracy of the information contained in this presentation, it is provided AS-IS without warranty of any kind, express or implied. IBM shall not be responsible for any damages arising out of the use of, or otherwise related to, this presentation or any other materials. Nothing contained in this presentation is intended to, nor shall have the effect of, creating any warranties or representations from IBM or its suppliers or licensors, or altering the terms and conditions of the applicable license agreement governing the use of IBM software. All customer examples described are presented as illustrations of how those customers have used IBM products and the results they may have achieved. Actual environmental costs and performance characteristics may vary by customer. Nothing contained in these materials is intended to, nor shall have the effect of, stating or implying that any activities undertaken by you will result in any specific sales, revenue growth or other results. © Copyright IBM Corporation 2014. All rights reserved. — U.S. Government Users Restricted Rights – Use, duplication or disclosure restricted by GSA ADP Schedule Contract with IBM Corp. — Please update paragraph below for the particular product or family brand trademarks you mention such as WebSphere, DB2,Maximo, Clearcase, Lotus, etc IBM, the IBM logo, ibm.com, [IBM Brand, if trademarked], and [IBM Product, if trademarked] are trademarks or registered trademarks of International Business Machines Corporation in the United States, other countries, or both. If these and other IBM trademarked terms are marked on their first occurrence in this information with a trademark symbol (® or TM), these symbols indicate U.S. registered or common law trademarks owned by IBM at the time this information was published. Such trademarks may also be registered or common law trademarks in other countries. A current list of IBM trademarks is available on the Web at •“Copyright and trademark information” at www.ibm.com/legal/copytrade.shtml •If you have mentioned trademarks that are not from IBM, please update and add the following lines:[Insert any special 3rd party trademark names/attributions here] •Other company, product, or service names may be trademarks or service marks of others. 42 We Value Your Feedback! • Don’t forget to submit your Insight session and speaker feedback! Your feedback is very important to us – we use it to continually improve the conference. • Access the Insight Conference Connect tool to quickly submit your surveys from your smartphone, laptop or conference kiosk. 43 Thank You