At the beginning of 2019, it’s worth taking stock of IT Ops initiatives. One of the biggest takeaways from 2018 is that cloud adoption and the digitization of customer experiences are mainstreaming and it’s high time for IT Ops teams to catch up. A follow-on takeaway is that gaining visibility into these new domains is challenging. But is it possible to actually go backwards? Sadly, yes.
Can A Move to the Cloud Be Bad News?
This past year, one of the large synthetic APM vendors announced that they were deprecating their “legacy, physical synthetic-monitoring locations” and that they would be “replaced with 100% cloud-based synthetic-monitoring locations.” They have since rolled out that change. Sounds great, right? Unfortunately, it isn’t. By moving all monitoring vantage points into the public cloud, this change is deprecating sound visibility into digital experience delivery. To understand why, let’s look at what constitutes “good” digital experience visibility and then compare cloud-only vantage points to that standard.
Where You Monitor From Matters
Sound digital experience monitoring (DEM) requires the right vantage points so you can understand how DNS, CDN, hosting and third party API provider components are performing. Vantage points need to be representative of your customers, so here are three considerations to keep in mind.
Do you have sufficient geographical scope?
If you’re serving markets in a variety of countries and cities, you’ll want to measure digital experience from Internet connectivity in all those places. For example, ThousandEyes has thousands of Cloud Agent vantage points pre-deployed in 175 cities and 54 countries that you can run monitoring tests from.
Are you monitoring from the right ISPs?
Measuring from Tier 1 ISP vantage points sounds like the right thing to do, but B2B and B2C users don’t sit at the most central places of the global Internet infrastructure. Instead, they connect to business class and consumer class ISPs, so you need vantage points that connect to those types of providers. For example, ThousandEyes Cloud Agents connect to Tier 2 and Tier 3 business-class ISPs or last-mile providers.
What about Public Cloud providers?
To state the (hopefully) obvious, no end users sit in AWS, Microsoft Azure or Google Cloud data centers. Furthermore, the big three cloud providers have truly massive internal connectivity that makes them ill-suited to represent end users. Moreover, if you’re hosting workloads in the public cloud, then monitoring from the public cloud is even less indicative of end users because our Public Cloud Performance Benchmark research has shown that traffic going between public cloud providers never leaves their mutual backbones or direct peering interconnections. That doesn’t mean that vantage points in Public Cloud providers aren’t useful at all for end-user experience, but it does mean that they are severely limited in their relevance.
So, are public-cloud vantage points useful at all? Absolutely! Consider using public-cloud vantage points to monitor intra-region, inter-region, multi-cloud and cloud region to third party back-end API performance and connectivity. All of those aspects are important to your overall digital experience delivery and worthy of continuous measurement. However, if you or a vendor moves to exclusively monitoring from public cloud data centers, that is a significant loss of visibility.
Getting to Why is Critical
Great digital experience monitoring goes beyond just getting end-to-end application layer metrics from the right vantage points. If something goes wrong, your operations team needs to know why. Depth of visibility matters. That’s why you need to have Network Intelligence that gives your teams a lens into how every network including your WAN, the Internet and cloud provider networks are impacting app delivery. You need the ability to rapidly see correlated end-to-end network metrics and history, then drill down to correlated Path Visualization with hop-by-hop metrics plus rDNS and BGP ASN details, ISP traffic outages and visualizations of BGP routing changes. Without this level of detail, your Ops team can get stuck in finger pointing and in an endless search for answers. With this level of detail, your Ops team can gather compelling evidence of problem domain issues and escalate effectively internally or to external providers.
Let’s just explore another aspect of getting to “why” when things go wrong, which is that even public cloud providers sometimes have issues. Sometimes those are internal to them, and sometimes they are due to forces beyond anyone’s control like power outages or munged Internet routing halfway around the world. If your workloads *and* your monitoring are both in the same public cloud provider, when there’s an issue with that provider, you’re going to be both down and totally blind. Not a great plan for operational visibility.
Getting to why can make the difference between hours or days of service disruption or keeping your digital users and customers delighted. Sadly, traditional DEM doesn’t offer this depth of visibility and that leaves a painful gap for organizations trying to understand and improve their digital experience delivery. That’s why ThousandEyes has put so many years of work into creating (patented) cross-correlation from the app delivery layer, to network and path layers, to collective intelligence ISP traffic outage detection, and down to rich BGP visualization.
Move Your DEM Visibility Forward in 2019
If digital experience is important to your business planning in 2019, then you need to avoid getting your visibility deprecated by vendor moves. Don’t move backwards or stay static with DEM that doesn’t get you to the why. Learn more about Network Intelligence by reading our 2018 Network Intelligence Planning Guide. If you know that you need this visibility today, contact us and we’ll start a consultative process with your team to understand your needs and explore how Network Intelligence from ThousandEyes can help your organization.