A global survey of 505 networking, operations, cloud and architecture professionals suggests more IT teams than ever (80%) are now coping with network blind spots created by lack of visibility into cloud and Internet services.
Conducted by Dimensional Research on behalf of Broadcom, the survey finds a full 84% regularly learn about network issues from end users rather than IT teams, with more than three quarters (76%) reporting missing or slow access to data impedes issue resolution times.
A full 95% don’t get all the information they require from either their Internet service provider (ISP) or cloud service provider (CSP), the survey finds. Specifically, survey respondents are looking for path latency and information about node/ hop issues. Tied at 50% each, the next most critical insights are route changes (48%), distributed denial of service (DDoS) attack locations (47%), DNS issues (46%), historical performance by path (44%) and path packet loss (41%).
More than three quarters of respondents (78%) said network complexity has only increased in the last year. The mission now is providing the visibility needed to ensure the quality application experience regardless of how complex the networking environment has become.
Jeremy Rossbach, chief technical evangelist for NetOps at Broadcom, said as more applications are deployed outside the traditional four walls of a data center, IT teams are losing visibility into distributed applications. Most IT teams appear to lack access to any type of network tap that enables them to collect telemetry data that could provide that visibility.
Additionally, finding the expertise required to manage networks has become a major challenge. A total of 41% of respondents report that acquiring and retaining skills is a challenge, with 65% noting they are to varying degrees dependent on third-parties to manage their networks. Much of the networking expertise that was once available has been reduced because so many IT professionals have opted to specialize in other areas such as artificial intelligence (AI), noted Rossbach.
Regaining that visibility is critical as end users become less tethered to any one network. In fact, it’s already not uncommon for end users accessing an application in New York City to be benefitting from proximity to data centers, while other end users in more remote locations are likely to be complaining about how slow that same application is running for them.
As organizations start to better understand that application experiences are being determined by the amount of latency experiences across a global hodgepodge of networks, there is an increased awareness of a growing need for deeper levels of observability. Many organizations are now embracing open-source agent software, dubbed OpenTelemetry, to affordably instrument their applications in a way that helps provide that observability, but those efforts only go so far. IT teams need to be able determine which applications are being adversely impacted by a network service being provided by a specific cloud service provider.
Ultimately, each IT team will need to determine for itself how best to regain visibility into their application environments. The only thing that is certain is that without it, it’s only a matter of time before what today may seem like a relatively minor inconvenience becomes a major outage crisis.
