How HPE 3PAR Can Help YOur Mission Critical on Cloud : Seminar Protecting Mission-Critical Application Against Downtime
•Download as PPTX, PDF•
1 like•256 views
Andi Bahar's Presentation material during Seminar Protecting Mission-Critical Application Against Downtime at Grha Datacomm Jakarta, Thursday March 8 2018
Recommended
More Related Content
What's hot
What's hot (20)
Similar to How HPE 3PAR Can Help YOur Mission Critical on Cloud : Seminar Protecting Mission-Critical Application Against Downtime
Similar to How HPE 3PAR Can Help YOur Mission Critical on Cloud : Seminar Protecting Mission-Critical Application Against Downtime (20)
More from PT Datacomm Diangraha
More from PT Datacomm Diangraha (20)
Recently uploaded
Recently uploaded (20)
How HPE 3PAR Can Help YOur Mission Critical on Cloud : Seminar Protecting Mission-Critical Application Against Downtime
- 1. How HPE 3PAR can help your mission critical on Cloud Andi Bahar HPE Country Product Manager - Storage
- 2. Apps and data It’s driven by a new generation of apps and data New business models Improve operational efficiencies Increase productivity Enhance customer experiences Differentiated products and services New sources of data and insight (“things”)
- 3. But I&O Leaders Must Look ‘Beyond the Box’ Predictive Predictive Analytics to anticipate and prevent issues across the infrastructure stack Cloud Ready Ubiquitous scale with data mobility between systems, sites, and the public cloud Timeless Architectural and financial flexibility to go anywhere with futureproofed platforms and guaranteed availability For HPE and Channel Partner internal use only
- 4. Business Impact is Severe DataApps Loading… You can’t afford an app data gap1 For HPE and Channel Partner internal use only
- 5. There is no one cause… and it’s too complex for humans to fix Source: InfoSight analysis HPE customer base Cross-stack Best Practices 54% Non-Storage 46% Storage Cross-stack Resource Contention Cross-stack Interoperability Storage Related Storage Network Compute VM/Container/DB For HPE and Channel Partner internal use only
- 6. Overcoming Complexity and Eliminating Anxiety 54% Cases solved outside of storage* Support you actually like >99.9999% Measured availability* Global visibility and learning 86% Issues automatically opened and solved* Predicts and prevents problems For HPE and Channel Partner internal use only
- 7. HPE InfoSight Extends to 3PAR Foundation for Predictive Support Predict problems & automate resolution for 3PAR Cloud-based Visibility Advanced visualizations, system health, and predictive analytics VMvision Cross-stack analytics for virtualized environments First release includes:
- 8. Eliminating the Guesswork in Managing Infrastructure Preemptive Recommendations for Your Data Center Example: Change network setting at port 3 to avoid failover issue Example: Move VM3 to Host2 as Host1 is oversubscribed Example: Apply QoS to vol1 to improve performance on vol2 Prevent Issues Before They Occur Improve Performance Proactively Optimize Available Resources
- 9. Performance impact scores Correlated performance factors Clear recommendations Automated cases with resolution HPE InfoSight...Gives Answers IOPS Machine Learning Correlation Traditional Monitoring…Creates Questions Should I be concerned? What’s causing it? How do I fix it? Latency ! Prescriptive Insights Beyond Monitoring
- 10. Storage Network Compute VMs/DBs/Apps Cross-Stack Analytics for VMware Environments Latency Attribution Identify root cause across host, storage, or SAN Noisy Neighbor Determine if VMs are hogging resources from another VM Host & Memory Analytics Visibility into host CPU and memory metrics Top Performing VMs Visibility into Top 10 VMs by IOPs and Latency Inactive VMs Visibility into inactive VMs to repurpose/reclaim resources
- 11. Diagnoses Abnormal Latency with VMVision HPE InfoSight VMVision Latency(ms) 0 20 40 From ToJan 24, 2017 March 20, 2017 Monday, March 6, 2017 24 Jan 30 Jan 4 Feb 7 Feb 11 Feb 15 Feb 19 Feb 22 Feb 26 Feb 1 Mar 4 Mar 7 Mar 10 Mar 14 Mar 16 Mar Host Network Storage Latency Spike Latency Host 0.3 Network 0.13 Storage 0.24
- 12. Diagnoses Abnormal Latency with VMVision VDI Sharepoint Datastore: esxi289-1x I/O Total: 5,278,000 Avg Latency: 23.31msec Latency(ms) 0 20 40 From To Host Network Storage Latency Spike Latency Host 0.3 Network 0.13 Storage 0.24 Jan 24, 2017 March 20, 2017 Monday, March 6, 2017 24 Jan 30 Jan 4 Feb 7 Feb 11 Feb 15 Feb 19 Feb 22 Feb 26 Feb 1 Mar 4 Mar 7 Mar 10 Mar 14 Mar 16 Mar
- 13. Takes the Guess Work Out of Planning Predicts Upgrade Needs Sizes Infrastructure Based on Applications Application Oracle data (GB) Apply Compression Percent reads … Oracle Database 1600 Yes 50 AF5000 AF7000 AF1000 … Array Quantity Recommendation 1 1 2 … R Recommended R Recommended T Not Recommended … HPE InfoSight Answers: How has my data usage trended? When am I going to run out of capacity? What’s the right SKU that I need for Oracle? What if I ran these apps...on the same array? Likely to exceed threshold: Sept 25 !
- 14. Acceleration of established and emerging applications App Acceleration Application workload integration Databases Virtualization ContainersBlock / File Control data Data payload Inter-Node & Cache IO RAID Calculations Data Deduplication Space Reclamation Mixed Workload 3PAR ASIC Replication Snapshots Compression RAID Rebuilds Thin Provisioning Sub-Lun Tiering CPU Multi-tenant and parallelized IO processing Every controller can access every disk, port, and resource #1 for every workload in Gartner’s Critical Capabilities report 2 years in a row
- 15. HPE 3PAR data-centric architecture – I/O path flash optimizations Adaptive Sparing One technology to improve flash performance, capacity and endurance Express Layout Drives performance and efficiency to maximize utilization of SSDs Persistent Checksum Guarantee host-to-disk consistency, eliminating data corruption System-Wide Striping All resources are used evenly, from front-end ports through CPUs to SSDs 1 2 3 4 1 2 4 SolidStateDisks Hostservers 3 3PAR ASIC
- 16. Legacy platforms Non-disruptively* migrate existing workloads from legacy arrays onto HPE 3PAR using Online Import Non-disruptive migration and technology refreshes to 3PAR – Array-level tools to migrate data quickly and easily Single array or federated systems Legacy 3PAR Peer Motion 3rd Party Online Import 3PAR Federation Federate HPE 3PAR systems together to form a single pool of resources for management and provisioning Technology refresh Migrate workloads from existing 3PAR systems onto new 3PAR platforms with no disruption to services
- 17. HPE RMC – Combining the performance of replication with the protection of backups Fast Deliver on SLAs with fast, non-disruptive, application consistent snapshots, backup and recovery Efficient Reduce cost & complexity with backup direct from 3PAR to StoreOnce Simple Control backup & recovery direct and seamlessly from their native hypervisor and application interfaces Reliable Protect applications with the availability of snapshots and the protection of backups 23x Faster backup 15x Faster restore *compared to traditional server-based backup environments RMC 3PAR StoreOnce
- 18. Architected for Efficiency Today & for What’s Next Deduplication Compression Only HPE Data Packing Max system efficiency with near-zero garbage Zero Detect1 2 3 4 Guaranteed Data Compaction Always In-line and selective Ready Storage Class Memory HPE 3PAR Adaptive Data Reduction Optimize Investment 50% Lower latency 60% Increased IOPS 100% Future proof HPE 3PAR 3D Cache First with SCM HPE 3PAR 3D Cache An extension of DRAM cache based on Storage Class Memory (Intel® Optane™) over NVMe to deliver accelerate flash itself NEW: HPE 3PAR Flash Now - All-flash
- 19. 3PAR For HPE and Channel Partner internal use only
- 20. Thank You! 20
Editor's Notes
- What happens when flash is a given and everyone’s a ‘leader’?
- InfoSight delivers 3 primary benefits. It predicts and prevents problems across the stack. It provides customers the visibility they need to quickly resolve complex issues. And, it delivers a support experience that customers actually like . Let’s unpack this. 1st. Our goal’s not to just show you that you have a problem and how to resolve it resolve it, but to prevent you from having a problem to being with. As InfoSight analyzes the installed base, it’s predicting and preventing in every customers' environment. And if it uncovers one, it proactively resolves it. And these are problems that exist outside of storage even. For instance, we’ve prevented thousands of systems from experiencing issues across the network, servers, and hypervisor. 2nd. InfoSight sees what others can’t, providing you clear insights up and down the infrastructure stack, across your environments, in the past and into the future. Traditional tools cause more trouble than good – requiring youto be their own data scientist and interpret what the data means. Instead, Infosight has embedded data science and machine learning to just give the right answers. And with all the systems connected in the cloud, infrastructure gets smarter learning from each other. 3rd. Because of InfoSight, we’ve transformed te support experience...one that customers actually like. InfoSight automates the tasks handled by traditional level 1 and 2 support staff. This has allowed Nimble to build a support organization made up entirely of level 3 experts. So in the rare case that customers need support, in less than min customers speak directly with a Level 3 expert who will quickly resolve the problem, even if it’s outside of storage. Customers can forget the mundane questions and painful escalations. And the proofs in the numbers...we’re 6-9s across our installed base for all systems including the first array that shipped | 86% | 54%
- HPE InfoSight Tells What to Do to: (read slide) So our AI recommendations eliminate the guesswork in managing infrastructure by making decisions for IT that ensure the optimal environment all the time. This paves the path for an autonomous data center as today we tell our customers what to do and tomorrow we can just do it for them.
- InfoSight picks up where monitoring falls short. One of the big limitations with monitoring is that it just creates more questions (read bullets)… and looking at latency is misleading because different workloads carry different inherent amounts of latency, so you end up chasing after the wrong events and get false alarms… whereas InfoSight gives answers. It applies a machine learning algorithm that filters the noise and simply highlights problematic periods based on the underlining I/O signatures – and shows what’s causing it and how to resolve it.
- Now when it comes to performance diagnostics - you have to a complete view and be able to correlate storage with other infrastructure systems to diagnose problems. Take for instance latency - it’s easy identify that there’s excessive latency, but diagnosing the root cause – be it in the host, VMs, network, array – that’s the challenge. So we’ve built in cross-stack correlations made available through our customer portal and this gives customers a single source of truth, which is customers to…(read bullets)
- Take for instance this example here with VM latency. Traditional monitoring tools can show you that latency occurred on a time-series graph, but that doesn’t help you diagnose the problem so you can fix it. With VMvision, a feature in the InfoSight customer portal, we pull data from your VMware environment and correlate this with the data we have across the infrastructure to clearly tell customers whether the host, network, or storage array is causing the latency.
- But we don’t stop there. Here we’re able to see that the server is contributing the most to the latency and then drill even further to see that it is the VDI application holding it up. And based on the root cause, customers can improve their application performance moving to a different data store, applying QoS controls, optimizing resources, and upgrading components.
- InfoSight also takes the guess work out of planning. It accurately predicts future capacity, performance and bandwidth needs and right sizes your new infrastructure through app-centric modeling. Our forecasting models are statistically highly confident using autoregressive and monte carlo simulations. And, planning for new infrastructure is simple, because InfoSight lets customers model different scenarios, applications, users, and performance requirements using a machine learning model that continously improves.
- All of the infrastructure we talk about is simply there to help your business and much of that comes from the applications that power your innovation agenda. How can the right infrastructure support your application transformation? First: the underlying architecture - Build for speed… designed for IO, not for a specific Flash Media such as NAND Flash - Built for massive consolidation of mixed application loads… which is what Hybrid IT is - ASIC to offload processing of many core functions so the system can maintain quality of service Second: the application ecosystem - Mode 1 apps… oracle, HANA, Microsoft, VMware… deep application integration Example is VMware vVOLS technology… the picked HPE as the design partner and reference platform - Mode 2… Partners like Docker and Mesosphere which HPE actually is investing in a t a corporate level
- 3D Cache works as an extension of DRAM cache and combines HPE 3PAR’s intelligent caching algorithms with Storage Class Memory (in this case Intel® Optane™) and NVMe to deliver extreme application acceleration