We’ve recently been putting tons of energy into evaluating the overall market landscape around Internet Performance, Application Performance and Network Performance. Given the success of New Relic, a major Dyn customer, partner, vendor, and comparator, in their successful IPO, we wanted to share this post to highlight the evolution around our overall space and what the future holds. Gartner calls this new space ‘Next Generation Infrastructure Tools,’ and IDC calls it ‘Systems Infrastructure Software Software-as-a-Service (SIS SaaS).’ We’re excited to discuss and explore its long-term prospects.
In talking with Will Cappelli of Gartner, we’ve recently become obsessed with the development of the ‘IT Operations Analytics’ space, as it’s been a super hot area that is garnering lots of attention for both startups and enterprises alike. Friends like New Relic, Pager Duty, Cedexis, ThousandEyes, AppDynamics, Netscout, Catchpoint, and others are crushing it.
When digging a layer deeper into the relatively short history of the space, where enterprises started doing performance monitoring with predictable statistical models in place, companies were trying to prove or disprove these at the outset. How fast should my application be? How should this website perform? Expectations were set in stone in a mathematical model to then be analyzed for improvements. This data modeling, where the model represents the reality, was and still is in existence for the data to simply be there to back up said model.
In a new-school, data-first approach (of which New Relic is thriving), the model doesn’t matter nearly as much, as enterprises look to ingest as much data as possible and as easily as possible. Once they have it, they can decide what to do with these newly packaged and powerful analytics. The shift outlined then brings that data into these sophisticated machine learning platforms (algorithms or rules engines) with an eye on making all of that great data actionable with performance for customers supreme. The problem is that the full solution ends up being many systems and APIs cobbled together, where you get the data and analytics, but the action is up to you or another system.
The reason this becomes more and more important as the Internet becomes the de facto connectivity tissue for commerce, communication and relationships is because current Internet environments have never been more volatile (see Missouri DDoS, Russian Internet, UK Atomic Bombs, Earthquake in Nepal). We report and cover issues extensively that impact the performance of the Internet at our Dyn Research site and on Twitter at @DynResearch.
Any model, if valid from the outset, is only valid for a short period of time. Yes, this makes life very hard for technical operations staff. The idea that companies can organize the data, then act on the data without human intervention, is an evolution the IT operations industry has been waiting for. There is a shift from ‘model-centric’ to ‘data-centric’ happening that New Relic and others are helping make possible. These are exciting times for similar providers, and more so even for the customers of these providers. But, how do you enact ACTION on all that information? That comes next.
The real positive here is that these technology and tools have democratized the space and enabled a major breakthrough in capability for companies that don’t have the time or resources – or PhD talent in data science or mathematics. The vendors building and deploying solutions are focused on ease of use and exposing business value propositions. Not only that, but universities are creating an enormous amount of education (courses, majors, materials, content) to ensure a growing number of decisions.
We recently kicked this around internally after several calls with industry folks, most notably Will Cappelli from Gartner, and realized the evolution is happening at different paces in different market segments. Said another way, the cloud-native Startup or Web brands are there already in demand and adoption, where the traditional enterprise is playing serious catch up.
Check out the five stages of evolution in the ‘IT Operations Analytics’ space, where we believe everyone is in a race to stage five and beyond.
IT Operations Analytics Stages:
Emphasis on gathering data together and making easy to access, data warehouse for analytics (Splunk era)
Effective data aggregation into statistical metrics, summarizing and presenting aggregation, with powerful visualization, automated base lining, trending (Tableau era)
Machine learning, multi-variable, anomaly detection, predictive analytics and pattern discovery, sophisticated correlations, multi-data source (New Relic era)
Complex statistical models become action-based to correct any faults or prevent faults from happening, validation and performance corrections, intervene and change value of one variable to another. Fairly rich theories and rules engine/algorithms (Sumo Logic era)
It’s an interesting and necessary shift to a causal analytics model (easier to understand) from statistical models, which need you to have that data science PHD to understand. “The action will be a direct consequence of the analytics”, said one esteemed analyst, in a causal model.
What we’re addressing at Dyn is in fact the big gap in what a lot of enterprises are trying to see. How does the Internet (aka end-users) access my website, application or network? Is it consistent? Is it optimized? Is it highly available? Where should my next data center go? What transit provider should I use? Who is the best CDN in Asia Pacific or EMEA or S. America? How do I adopt the cloud? Should I be afraid of SaaS services? Am I losing control? Are we delivering the best experience for end-users? Always? Can you make changes, based on cause-and-effect correlations, on the fly?
We’re at the forefront of causal model approach, which helps bring the facts, data and technology rules engines together to inform ACTION. If we can execute on this vision, the future of the Internet and it’s performance will be very optimized for all. To learn more about our efforts, please peruse Dyn.com for more information and download our white paper about Dyn Internet Intelligence.
Kyle York is Dyn’s Chief Strategy Officer and has been a long-time executive, having joined in 2008. Over the years, he has held go-to-market leadership roles in worldwide sales, marketing, and services. In his role as CSO, Kyle focuses on overall corporate strategy, including: positioning and evangelism, new market entry, strategic alliances and partnerships, M&A, and business development. Outside of Dyn, Kyle is an angel investor, entrepreneur, and advisor in several startups. Follow Kyle on Twitter: @kyork20 and @Dyn.