How to Assure Performance in Hybrid Cloud Ecosystems

Editor's Notes

• #2: Welcome to today’s webinar --- Assuring Performance in Hybrid Cloud Ecosystems...   I’m Natlie Tomko and I’ll be moderating today’s session.
• #3: As usual let’s begin with a few logistics....   You can submit a question any time by using the questions panel on the go to webinar client....if we can’t get to your question you’ll get an answer by email after the session today...   ...we’ll be recording the session and all attendees will get both a link to the session and can get copies of the slides just by going to SlideShare at the link shown here....  
• #4: We have a couple of panelists today.... John Worthington is eG’s Director of Product Marketing and will be leading us through our discussion ....   ...and we’ll also be joined by Bala Vaidhanatham -- our Chief Technology Officer ---   as always if you have any specific questions just use the chat panel and we’ll be happy to address them!   So, with that as our introduction I’d like to give it to John to get us started....
• #5: Thanks Natalie! .... and welcome everyone...   Here’s our agenda...   We’ll begin with a quick reminder of the importance of quality of experience in the digital age...   ...then we’ll talk about the move to hybrid cloud ecosystems and why total performance visibility is so critical for enabling operating model change...   ....we’ll talk about what a digital service performance baseline is and how critical it is to cloud migrations.... ...and we’ll wrap it up by giving you some ideas about how you can accelerate your hybrid cloud journey and reduce your risks along the way....  
• #6: Most of us understand that the business is totally pre-occupied with user experience, and for good reason....   <click> Gartner and others have made it pretty clear that customer experience is the new basis for competition in the digital age...   ...and this makes quality-of-service and site/application performance business-critical....   ...the performance of your site and applications are now the foundation of end user experience in the digital world...   a link to a recent webinar where we talked about this is on the slide...
• #7: Here’s the familiar layered cloud service model...it’s an easy way to understand at a high level what YOU have to manage versus what your cloud provider manages in different cloud service models....   ...<click> of course, the number of cloud services--- and cloud service providers– has exploded.   So, we’re definitely in a hybrid, multi-cloud world....
• #8:   ...and it’s really important for customers to recognize that when it comes to digital service performance...   <CLICK> while you can let your cloud provider manage infrastructure using IaaS ...   <CLICK>and you might even be able to let them manage some or all of an application using PaaS or SaaS...   The <CLICK> digital business transactions and digital user experience that your customers use are still your responsibility, right?   <CLICK> so when it comes to your digital user experience, the buck still stops with YOU
• #9: This raises an important point...the cloud is not a place or a technology, even though many of us talk about it like it is.... public, private or hybrid cloud computing are operating models....   <CLICK> ...the familiar people, process, technology is a common example of the elements of an operating model...and it’s the operating model that is the organization’s “system” for delivering value to customers  
• #10: So, the reality of a hybrid cloud world is that there may be multiple operating models in play...   ... you could have legacy systems in a private data center, you could have private cloud services and you could have different forms of public cloud services...   ...<CLICK> and as you introduce public cloud services the cloud providers will being to introduce operating model change....   ...<CLICK> which can further complicate what may already be multiple internal operating models in play within your organization, whether that be ITIL, DevOps or something else...   ...and as we said earlier managing the performance of digital business transactions and digital user experience remains on YOU <CLICK>...so, a unified view of performance has become business-critical to competitive advantage, especially in the hybrid cloud world!....
• #11: So, the first part of our discussion today centers around this question...   ... How can unified, consistent views of performance enable effective service delivery?
• #12: Let’s quickly talk about these 5 ways you can achieve a consistent, unified view of performance ...   ...establishing and end-user view   ...providing service team members with an end-to-end and top-to-bottom view of performance   ...leveraging machine learning and analytics   ... and giving all stakeholders role-based views of performance and active dashboards where needed...
• #13: Both simulated and real user monitoring are needed for effective monitoring of the end user experience... <CLICK> client emulation can tell you whether the service is up, how fast it’s running and whether all transaction are running...whether there are end users using the applications or not.... <CLICK> real user monitoring can tell you how your users are interacting with the service, what the actual user experience is like, and where the performance issues are...that is, whether they’re in the front-end or the back-end...
• #14: Providing monitor users with an end-to-end view is also important...   <CLICK> this may include private topology views... <CLICK> hybrid cloud topology views or   <CLICK> views of cloud resources such as SaaS applications or public clouds, availability zones or instances... These visualizations should let you quickly see where along the end-to-end topology the source of performance bottlenecks are...
• #15: ...of course, once issues are identified in a particular tier, monitor users will need to get more information...   ...and this information may need to be quite specific depending on the technical domain <CLICK> such as traditional infrastructure components like servers, storage, network devices and so on....   <CLICK> virtual, converged and application components ....   <CLICK> as well as cloud-based components ....   ...each of these components will have layered dependencies...very similar to the layered service model we saw earlier...  
• #16: Machine learning and analytics are topics that everyone’s talking about....   ...from an eG Enterprise perspective, we use a Universal Agent or data collector that can collect data in an agent-based or agentless manner, using whatever methods are available...from perfmon and SNMP to APIs....   ...we’ve been using machine learning techniques for over a decade to automatically learn the norms of all measurements, and patented analytics to automatically correlate metrics based on the end-to-end and top-to-bottom dependencies of a digital business service... and isolate which layer of which component is the root-cause of a performance issue.   The accuracy and ease of use of our analytics has always been a hallmark of eG Enterprise....
• #17: ...and finally, monitor users in an enterprise have many different personas....today each user wants a personalized dashboard that is tailored to their role in whatever operating models exist in your organization...these fall into three basic types   ...executives will only be interested in high-level KPIs and will want these linked to business impacts...   ...operational users will want real time data, and perhaps aggregated views, views by service or by technical domain...   ...and analytical users will need active dashboards...that is, the ability to drill down and go directly to detailed diagnostic data...without having to window into another tool   ....with this as background I’d like to turn it over to Bala who’ll give us a quick look at what some of this looks like...then we’ll take a few questions Bala? DEMO OK, before we continue let’s take a few questions...we’ll take questions at the end too... SEED QUESTIONS See word doc
• #18: Thanks for those questions and if we can’t get to you we’ll respond via email so keep using the chat panel... now that we’ve seen an example of how unified performance monitoring can give stakeholders a consistent view of performance, let’s put this into a cloud migration context...
• #19: ...this slide is from last July’s webinar ... We’ll focus on the three activities at the bottom of each phase... <CLICK> getting an initial baseline, validating performance norms during migrations and using the baseline for ongoing improvement...
• #20: Of course there are different approaches to capturing a baseline, from spreadsheets to specialized modeling tools....and part of the challenge is that the most effective baseline will combine information from every level of every tier of a particular workload.... ...so, the bottom line is that the more comprehensive your baseline, the lower your migration risk and the greater the likelihood that your user experience will be equal to or better than it was before the migration <CLICK> you also need to do baselining at every stage of the service lifecycle, and in real time in production....and trying to do this with multiple tools can be extremely difficult <CLICK> this is one reason why machine learning is such a hot topic...manually setting thresholds is just no longer do-able... We’ve been using machine learning to automatically set time-varying thresholds for over a decade now, and every measurement and metric we capture is automatically baselined....
• #21: So, during pre-migration baselining we’ll want to understand user experience <CLICK>....how, when and where user experience is impacted....   <CLICK> understand user behavior based on activity and its impact on resources....   <CLICK> understand application usage...   <CLICK> peak utilization....and <CLICK> resource consumption during peak demand....   All these baselines --- at the user experience, application and infrastructure levels --- are important for planning an effective migration and minimizing migration risk.
• #22: <CLICK> Machine learning and auto-baselining is a critical enabler for automatically isolating the effects of changes during migration   <CLICK> and this is enhanced by the ability to automatically track changes, compare them over time and automatically correlate them to performance impacts...   <CLICK> and we'll also perform event analysis and trending during migrations so we can give you early warnings of performance issues before they get out of hand...
• #23: It’s important to recognize that hybrid cloud ecosystems will drive operating model change...<CLICK> and most organizations will have multiple operating modes in play...   ...people are the most difficult part of operating model change and this is why both ITIL and DevOps talk about ongoing cycles of feedback and communication...   ... total performance visibility can enable these feedback loops, improve organizational transparency and trust and accelerate cultural change.
• #24: ...before we wrap it up today let’s offer a few comments on the best way to get started on a hybrid cloud journey....
• #25: This generic path shows high-level activities in the Pre-Migration, Migration and Post-Migration phases of a cloud migration. Of course a cloud migration will include many activities besides monitoring. So, it’s common that you’d see other workstreams like financial management and security... ... but because of the operational nature of monitoring and its broad technical and organizational scope we sometimes delay it or try to fragment it into other workstreams.... So sometimes you may not see performance monitoring as a specific program of work within the transformation path.... ...and this tends to perpetuate the status quo from a monitoring perspective and can result in an ineffective baseline of performance... ...this significantly increases the risks associated with any cloud migration. So, try to create a program of work around performance monitoring...
• #26: ...this may seem obvious, but there’s a natural tendency to try and skip steps or not be honest about where we’re starting from... any transformational change program---whether you call it cloud, DevOps or something else -- everyone’s starting point tends to be different. In a webinar we had last year, we polled the audience from a monitoring perspective and found that more than 80% of them were still not at a services oriented approach to their monitoring or tended to be fairly reactive... So, just be honest about your starting point....are you able to get a comprehensive baseline of end-to-end and top-to-bottom performance of your current IT services? .... that's really what you'll need to assure the performance of your hybrid cloud ecosystems...
• #27: ...from a technology perspective, a lot of organizations are experimenting with devops, containers and microservices.... ...but almost all also have packaged software applications or legacy applications that won't be modernized for the cloud for some time, even if the workloads happen to run in a cloud provider's data center... ...the reality for almost all enterprises is going to be a mix of cloud-based, virtual and legacy ecosystems...be sure your monitoring approach takes these realities into account as well
• #28: ...and finally, remember no matter what mix of cloud services you use, your customers are not going to care who is managing which part of your hybrid cloud ecosystem <CLICK>....   ....when it comes to digital business transactions and the digital user experience, your customers are going to look to YOU not your providers! <CLICK>  
• #29: eG can work directly with you or through our network of Authorized Partners and help you along your journey to the hybrid cloud... ...of course our focus is on monitoring, and we can help you manage performance monitoring as a program of work and as a critical element of your hybrid cloud migration... this could include training and knowledge transfer, performance baselining as well as services-oriented measurement and reporting services... contact your eG representative to learn more... using this approach we can help you establish cycles of monitoring that get the most out of your investment
• #30: OK, let’s see if we have any more questions....