Optimizing the Network for Reliable Application Delivery Across the Cloud
1 like523 views
The document discusses strategies for optimizing network performance for application delivery in cloud environments, with an emphasis on survey results regarding cloud adoption, performance issues, and optimization methods. It defines various cloud models (private, public, hybrid) and outlines key findings related to cloud workloads, monitoring, and the effectiveness of traditional and next-generation optimization solutions. The document highlights the significance of pre-deployment testing, performance issues experienced, and the optimization techniques utilized to enhance cloud application performance.
Optimizing the Network for Reliable Application Delivery Across the Cloud
- 1. Optimizing the Network for Reliable Application Delivery Across the Cloud August 5, 2015 Tracy Corbo @tcorbo Principal Research Analyst Network Management Enterprise Management Associates
- 2. Today’s Presenters Slide 2 © 2015 Enterprise Management Associates, Inc. Tracy Corbo Principal Research Analyst Enterprise Management Associates (EMA) Tracy has 24 years of experience in the computing industry, both as an industry analyst and on the vendor side doing competitive analysis. At EMA, Tracy covers Network Management. Prior to joining EMA, Tracy worked for two of the largest IT market research vendors, Gartner and IDC.
- 3. Logistics for Today’s Webinar Slide 3 © 2015 Enterprise Management Associates, Inc. An archived version of the event recording will be available at www.enterprisemanagement.com • Log questions in the Q&A panel located on the lower right corner of your screen • Questions will be addressed during the Q&A session of the event Questions Event recording
- 4. Optimizing the Network for Reliable Application Delivery Across the Cloud August 5, 2015 Tracy Corbo @tcorbo Principal Research Analyst Network Management Enterprise Management Associates
- 5. Sponsors Slide 5 © 2015 Enterprise Management Associates, Inc.
- 6. Agenda • Defining cloud • Survey objectives • Demographics • Cloud adoption, connectivity, and monitoring • Cloud performance issues • Optimization choices • Satisfaction • Drivers and benefits • EMA Key Takeaways Slide 6 © 2015 Enterprise Management Associates, Inc.
- 7. Agenda • Survey objectives & demographics • Cloud adoption, workloads, connectivity, and monitoring • Pre-deployment testing • Optimization choices • WAN optimization • Application optimization • Alternative WAN optimization • Managed Internet providers • Switching to dedicated link • Satisfaction • Drivers and benefits • EMA Key Takeaways Slide 7 © 2015 Enterprise Management Associates, Inc.
- 8. Defining Cloud For purpose of this survey, EMA defined cloud as follows: Slide 8 © 2015 Enterprise Management Associates, Inc. Private Cloud – The private cloud infrastructure operates solely for a single organization and is managed and hosted internally, inside the corporate firewall. Public Cloud – The public cloud infrastructure is provisioned for shared use. It may be owned, managed, and operated by a business, an academic or government organization, or some combination of those organizations. The cloud service exists on the premises of the cloud provider, outside the corporate firewall. Public cloud includes services such as Amazon Web Services (AWS), Microsoft Azure, IBM/SoftLayer, and Google Compute Engine. Hybrid Cloud – The hybrid cloud infrastructure is a combination of public cloud services and an on-premises private cloud. These systems are tied together via APIs or other proprietary technology that enables data and application portability between the two systems, but they remain separate entities. For example, sensitive data may remain on the private cloud, but the applications or services reside on an external public cloud.
- 9. Survey – Primary Objectives • Do performance issues exist across all three types of cloud deployments? • How are performance issues being resolved across the various types of cloud? • Do traditional optimization methods have a role to play in cloud? • Are next-generation cloud-based optimization solutions gaining traction? Slide 9 © 2015 Enterprise Management Associates, Inc.
- 10. Survey – Secondary Objectives • What workloads are being deployed across the different cloud models? • Are those workloads changing? • How are cloud workloads being monitored? Slide 10 © 2015 Enterprise Management Associates, Inc.
- 11. Demographics Slide 11 © 2015 Enterprise Management Associates, Inc. CIO/CTO (IT Executive Management) 31% IT-related Director (or Equivalent) 25% IT-related Manager/Supervisor (or Equivalent) 11% IT-related Project/Program Manager 7% IT-related Architect 5% IT-related Consultant/Integrato r 4% IT-related Business Analyst 4% Other 13% Which of the following best describes your role in the organization? N= 152
- 12. Cloud Adoption
- 13. Penetration of Cloud Computing in Enterprises Slide 13 © 2015 Enterprise Management Associates, Inc. 38% 35% 9% 5% 8% 5% It is currently adopted and is an essential part of our business. It is currently adopted and an important part of our business. We are currently generating revenue through delivered cloud services. We are or plan to generate revenue through delivered cloud services. It is currently adopted and mostly supplemental to other types of computing. We are still in the early phase of adoption. How would you describe the extent to which cloud computing has been adopted within your business/org?
- 14. Cloud Computing Essential to Small Enterprise Slide 14 © 2015 Enterprise Management Associates, Inc. 47% 36% 5% 4% 5% 2% 30% 45% 11% 0% 7% 7% 34% 25% 11% 11% 11% 8% It is currently adopted and is an essential part of our business. It is currently adopted and an important part of our business. We are currently generating revenue through delivered cloud services. We are or plan to generate revenue through delivered cloud services. It is currently adopted and mostly supplemental to other types of computing. We are still in the early phase of adoption. How would you describe the extent to which cloud computing has been adopted within your business/org? (by company size) Small Enterprise (500 to 2,499) Medium Enterprise (2,500 to 9,999) Large Enterprise (10,000 or>)
- 15. External SaaS Most Widely Adopted Slide 15 © 2015 Enterprise Management Associates, Inc. 69% 46% 45% 43% 42% 41% 39% 38% 34% 30% External SaaS-delivered apps Externally hosted IaaS Internally hosted Virt Apps Internal on-demand Virt Infra Internally hosted VDI (External) Web hosting via CSPs Internal IaaS Externally hosted PaaS Internal PaaS (External) App perf enhancemnts via CSPs What cloud-related services is your company leveraging?
- 16. Government Big Consumer of SaaS Cloud Services Slide 16 © 2015 Enterprise Management Associates, Inc. 90% 60% 60% 57% 52% 50% 50% 50% 48% 44% External hosted SaaS Internally hosted Virt App Internally hosted VDI Externally hosted IaaS Internal on-demand virt infra Internal IaaS capabilities Externally hosted PaaS capabilities Internal PaaS (External) Web hosting via CSPs (External) App perf enhance via CSPs Gov Fin/Ba nk/Ins Fin/Ba nk/Ins Prof Serv Fin/Ba nk/Ins Retail/ Whole /Distri b& Gov Retail/ Whole /Distri b Prof Serv Fin/Ba nk/Ins Fin/Ba nk/Ins Most popular cloud services by vertical
- 17. Cloud Workloads
- 18. Workloads Differ by Cloud Type Slide 18 © 2015 Enterprise Management Associates, Inc. 49% 48% 43% 37% 36% 35% 29% 29% 28% 0% 10% 20% 30% 40% 50% 60% Top Three Workloads by Cloud Type Public Hybrid Private
- 19. External Cloud Workloads Differ by Company Size Slide 19 © 2015 Enterprise Management Associates, Inc. 41% 20% 27%27% 43% 35% 30% 17% 39% Email Cloud-based Anti- spam/Anti-virus eCommerce Top Public Cloud Workloads (by company size) Small Enterprise Medium Enterprise Large Enterprise 29% 29% 16% 8% 30% 27% 43% 27% 30% 22% 30% 35% Corporate Website Office productivity BI/Big Data Finance/Accounting Top Hybrid Cloud Workloads (by company size) Small Enterprise Medium Enterprise Large Enterprise
- 20. Private Cloud Workloads Flat or Declining Slide 20 © 2015 Enterprise Management Associates, Inc. 43% 36% 35% 38% 28% 34% 34% 48% 25% 49% 31% 28% 40% 33% 32% 33% 29% 31% 33% 35% 24% 36% 30% 31% Private Cloud – Current vs. Planned Private Current Private Planned
- 21. Public Cloud Workloads Relatively Flat Slide 21 © 2015 Enterprise Management Associates, Inc. 19% 35% 37% 24% 26% 27% 23% 21% 34% 24% 36% 25% 18% 32% 33% 25% 23% 29% 24% 22% 34% 28% 31% 26% Public Cloud – Current vs. Planned Public Current Public Planned
- 22. Workloads Moving to Hybrid Cloud Slide 22 © 2015 Enterprise Management Associates, Inc. 22% 21% 23% 26% 23% 29% 28% 16% 22% 24% 29% 26% 16% 26% 27% 22% 26% 28% 29% 23% 25% 28% 29% 26% Hybrid Cloud Deployments – Current vs. Planned Current Planned
- 24. Mobility on the Rise in the Enterprise Slide 24 © 2015 Enterprise Management Associates, Inc. 39% 34% 37% 31% 26% 25% 25% 22% 24% 21% 24% 22% 20% 15% 39% 33% 24% 26% 29% 29% 26% 29% 27% 25% 21% 18% 17% 20% Wi-Fi Public InternetBroadband w/ VPN Cloud Exchange provider VPN/WAN extension Managed Business Internet Azure Express Route Public InternetBroadband, no VPN Dedicated 10MB fiber Direct connect from colo or hosting provider Carrier Ethernet/VPLS Carrier MPLS/Layer 3 MPLS VPN AWS Direct Connect VXLAN GRE Tunnels What types of network connectivity are in use for reaching your external cloud services? Public Hybrid
- 25. Biggest Mobility Impact in Large Enterprise Slide 25 © 2015 Enterprise Management Associates, Inc. 46% 24% 43% 34% 31% 51% Small Medium Large WiFi Connectivity (by company size) Public Hybrid
- 26. Cloud Monitoring
- 27. Private Existing – Public Reports from CPs Slide 27 © 2015 Enterprise Management Associates, Inc. 46% 30% 28% 33% 14% 43% 45% 31% 18% 16% 38% 34% 26% 24% 19% Use existing monitoring solution Rely on reports from cloud providers Purchased new monitoring solution for cloud Wrote custom code to accommodate cloud No plans for monitoring How do you currently monitor various cloud deployments? Private Public Hybrid
- 29. Performance Testing Prior to Deployment? – Yes Slide 29 © 2015 Enterprise Management Associates, Inc. 57% 50% 55% 53% 52% 19% 5% 2% 57% 57% 45% 49% 48% 24% 5% 4% 44% 47% 47% 43% 38% 20% 7% 7% Bandwidth Connectivity integrity Quality of Service (QoS) Stress/load testing Latency tests No pre- deployment… No tests conducted Don't know Did you conduct any of the following pre-deployment tests prior to deploying to the cloud? Private Public Hybrid
- 30. Network Connectivity Adjustments Made Slide 30 © 2015 Enterprise Management Associates, Inc. 58% 25% 34%33% 46% 39%40% 29% 47% Yes added bandwidth Yes changed connectivity type No, we made no changes For each of your cloud deployments, please indicate whether or not you had to make any adjustments to your network connectivity prior to going into a live production environment. Private Public Hybrid
- 31. Performance Issues Exist Slide 31 © 2015 Enterprise Management Associates, Inc. 53% 49% 52% 47% 51% 48% Private Public Hybrid Since deploying to the cloud, have you experienced any performance issues that are impacting end users enough to result in generating help desk tickets? Yes No
- 32. Private Cloud Performance a Major Pain Point for Large Enterprises Slide 32 © 2015 Enterprise Management Associates, Inc. 44% 42% 70% 56% 58% 30% Small Enterprise Medium Enterprise Large Enterprise Private Cloud Deployment Issues (by company size) Hybrid 62%
- 34. Traditional Methods Remain Relevant Slide 34 © 2015 Enterprise Management Associates, Inc. 47% 46% 40% 35% 32% 33% 27% 28% 10% 32% 29% 37% 37% 34% 22% 22% 28% 11% 41% 44% 34% 38% 28% 31% 31% 16% 12% Virtualized WAN optimization in hosted cloud Virtualized ADC/Application load balancing in hosted cloud WAN optimization on premises, virtual or physical WAN optimization option/add-on from cloud provider ADC/Application load balancing on premises, virtual or physical ADC/Application load balancing option/add-on from cloud provider Alternative WAN Optimization (e.g., WAN overlay, SD WAN, Hybrid WAN, etc.) Managed Internet (e.g., Akamai, Level3, etc.) Switched from Internet to dedicated link What types of optimization actions have you taken to improve cloud performance? Private Public Hybrid
- 35. Optimizing Network and Application Load Slide 35 © 2015 Enterprise Management Associates, Inc. 47% 46% 37% 37% 44% 41% Virtualized WAN opt in hosted cloud Virtualized ADC/App load bal in hosted cloud WAN opt on premises, virtual/physical WAN opt/add-on from cloud provider Virtualized ADC/App load bal in hosted cloud Virtualized WAN opt in hosted cloud PrivatePublicHybrid Top Two Optimization Choices by Cloud Type
- 36. WAN Optimization
- 37. QoS Least Utilized Slide 37 © 2015 Enterprise Management Associates, Inc. 41% 39% 46% 44% 32% 34% 39% 24% 28% 38% 45% 47% 45% 33% 36% 31% 22% 24% 53% 44% 33% 36% 40% 33% 31% 35% 24% Route optimization Caching TCP acceleration Data deduplication Compression Protocol optimization Traffic shaping CIFS acceleration QoS WAN Optimization Techniques Implemented Private Public Hybrid
- 38. Methods Vary by Cloud Type & Workloads Slide 38 © 2015 Enterprise Management Associates, Inc. 52% 53% 48% 59% 47% 50% 57% 65% 43% Data dedup TCP acceleration Traffic shaping Data dedup Data dedup TCP acceleration Route optimization Route optimization Caching SmallMedLargeSmallMedLargeSmallMedLarge PrivatePublicHybrid Top WAN Optimization Methods by cloud type & company size
- 39. 17% 12% 16% 18% 17% 12% 6% 1% 21% 24% 12% 14% 12% 9% 5% 3% 22% 15% 15% 9% 9% 13% 16% 2% Best fit for this type of workload Best fit within our budget Already using the product Easy to integrate with our backend systems Most cloud-friendly Easy to use Most feature-rich Free What was the primary driver for selecting this particular WAN optimization solution? Private Public Hybrid Reasons for Choosing WAN Optimization Slide 39 © 2015 Enterprise Management Associates, Inc.
- 41. Data Compression & SSL Acceleration Top Choices Slide 41 © 2015 Enterprise Management Associates, Inc. 44% 35% 39% 35% 32% 30% 38% 27% 32% 27% 15% 38% 43% 38% 28% 32% 25% 25% 32% 32% 21% 25% 49% 37% 33% 41% 33% 39% 29% 31% 20% 31% 24% Data compression SSL acceleration Content routing Content caching Layer 4 load bal Application acceleration Global server load bal Core protocol opt Layer 7 load bal SSL offload Round robin Are you implementing any of the following ADC-based or load balancing technologies? Private Public Hybrid
- 42. Most Cloud-Friendly & Cost Slide 42 © 2015 Enterprise Management Associates, Inc. 21% 17% 14% 20% 14% 6% 7% 1% 25% 15% 21% 11% 6% 6% 11% 6% 16% 22% 12% 16% 10% 14% 6% 2% Most cloud-friendly Best fit within our budget Best fit for this type of workload Easy to integrate with our backend systems Easy to use Already using the product Most feature-rich Free What was the primary driver for selecting this particular ADC/Load Balancer solution? Private Public Hybrid
- 44. CloudGenix Most Popular Across All Cloud Types Slide 44 © 2015 Enterprise Management Associates, Inc. 48% 44% 40% 24% 28% 24% 16% 4% 53% 47% 41% 47% 35% 12% 24% 0% 52% 33% 38% 29% 33% 43% 38% 5% CloudGenix Virtela VeloCloud Talari Networks Glue Networks Aryaka Viptela Other (specify): Which of the following alternative-vendor WAN optimization solutions are you implementing? Private Public Hybrid
- 45. Primary Driver Differs by Cloud Type Slide 45 © 2015 Enterprise Management Associates, Inc. 12% 16% 20% 16% 12% 16% 4% 29% 29% 18% 6% 12% 0% 6% 14% 14% 10% 14% 5% 19% 19% Easy to use Most cloud-friendly Best fit within our budget Best fit for this type of workload Most feature-rich Easy to integrate with our backend systems Already using the product What was the primary driver for selection this alternative WAN optimization solution? Private Public Hybrid
- 47. Slide 47 © 2015 Enterprise Management Associates, Inc. 58% 35% 35% 23% 12% 4% 50% 32% 36% 32% 27% 0% 64% 64% 27% 18% 27% 0% Amazon Cloudfront Limelight Akamai Cachefly Aryaka CDN-as-a- Service Other (specify): Which of the following managed Internet services are you implementing? Private Public Hybrid Amazon Cloudfront Most Popular Across All Cloud Types
- 48. Best Fit for Workload and Budget Primary Drivers Slide 48 © 2015 Enterprise Management Associates, Inc. 23% 19% 12% 12% 4% 19% 8% 4% 18% 9% 18% 14% 14% 9% 9% 9% 18% 27% 18% 18% 18% 0% 0% 0% Best fit for this type of workload Best fit within our budget Most cloud-friendly Most feature-rich Easy to use Easy to integrate with our backend systems Already using the product Free What was the primary driver for selecting this alternative managed Internet solution? Private Public Hybrid
- 49. Switching to Dedicated Link
- 50. Security & Control Top Drivers for Switching to Dedicated Link Slide 50 © 2015 Enterprise Management Associates, Inc. 56% 22% 22% 33% 44% 22% 50% 12% 38% Security Better end-to-end control over performance Best fit for this type of workload What was the primary driver for selecting a dedicated link? Private Public Hybrid
- 51. Satisfaction
- 52. Application Optimization Solutions – Highest Level of Satisfaction Slide 52 © 2015 Enterprise Management Associates, Inc. 21% 63% 15% 0% 23% 58% 17% 2% 16% 67% 16% 0% Excellent Very good Neither good nor bad Not good Application Optimization – Highest Level of Satisfaction Across All Three Cloud Types Private Public Hybrid
- 53. Drivers & Benefits Realized
- 54. On-Demand Scalability Tops List Slide 54 © 2015 Enterprise Management Associates, Inc. 14% 13% 13% 10% 9% 9% 8% 7% 7% 6% 5% On-demand scalability/elasticity Functional extensibility OpEx cost savings CapEx cost savings Reduce admin overhead Easy upgrades Pay per use Accelerated deployment Accelerated app/dev Internal cloud mandate Self service provisioning What was your primary driver for moving workloads to the cloud?
- 55. Easy Upgrades Most Realized Benefit – Not Primary Driver Slide 55 © 2015 Enterprise Management Associates, Inc. 72% 66% 59% 61% 59% 58% 56% 54% 53% 51% 51% 28% 34% 41% 39% 41% 42% 44% 46% 47% 49% 49% Easy upgrades On-demnd scalability/elasticity Accelerated deployment Functional extensibility Reduce admin overhead Internal cloud mandate OpEx cost savings CapEx cost savings Self service provisioning Pay per use Accelerated app/dev Since deploying to the cloud, which of the following benefits have been met and which remain to be seen? Achieved Not Yet
- 56. EMA Key Takeaways – Primary • Despite pre-deployment testing, performance issues persist. • There’s no “one size fits all” optimization solution. • Workloads and cloud type matter. • Traditional optimization technologies remain relevant, but next gen cloud based are making headway. • “Feature-rich” is not important — “Cloud-friendly” & “Best fit for workload & budget” matter more. • Cloud suffers from same monitoring, performance, and manageability concerns as traditional deployments. Slide 56 © 2015 Enterprise Management Associates, Inc.
- 57. EMA Key Takeaways – Secondary • Hybrid cloud workloads are increasing. • Proprietary connectivity is giving way to less-expensive public/broadband. • Fixed networks are giving way to wireless ones. • Mobile is the new enterprise desktop. • Pre-deployment testing should include mobile devices and wireless networks. Slide 57 © 2015 Enterprise Management Associates, Inc.
- 58. Learn more & access report at www.enterprisemanagement.com Slide 58 © 2015 Enterprise Management Associates, Inc.