When it comes to network downtime in any industry, the underlying statistics tell us one thing: it costs money and lots of it.
According to the Ponemon Institute, downtime can cost around an astounding $8 000 per minute, which means an hour-long network outage can cost a business over half a million dollars. A separate IHS report underlines this claim, saying medium and large companies that typically experience five minutes of downtime each month are looking at losses equating to $1 million and $60 million per year, respectively.
“Financial impact aside, a sluggish network will also affect IT operations resulting in slow backups and corrupt data as a result of dropped packets,” says Anton Jacobsz, managing director at Networks Unlimited. “In addition, business continuity and system availability can be compromised and all of this can result in critical customer-facing applications being affected, which, in turn, negatively affects your business brand and customer loyalty.”
A white paper titled Real-time Network Visibility from Mellanox, says that the symptoms of a poorly performing network are easy to see – but pinpointing the underlying causes of network issues is never quite as straightforward. It says the size, complexity, speed and dynamic nature of network traffic can make it difficult to identify bottlenecks and other issues.
“As our partner in network visibility, Mellanox’ paper underscores one message we’ve been driving home for quite some time and that is, you can’t manage what you can’t measure,” says Jacobsz. “IT security and analytics are only as good as the data they see, and it is IT’s core challenge to ensure that the infrastructure supporting these tools delivers reliable, fast and secure applications. In order to achieve this, IT needs total visibility into the network at all times.”
Businesses struggling with lethargic networks might want to look at their network on a very granular level, drilling down to switch level, he explains, because it is at switch level that the network build began.
My network for a switch?
Only those operating deep within the IT administration truly understand exactly how choosing the right switch can heighten visibility, but it is vital that people across all levels of the business have at least a broad understanding of their impact, says Jacobsz. “Network telemetry provides the link between the network and business-critical application performance, ensuring visibility into real-time information.”
Businesses using the right Ethernet switches coupled with a solid analytics tool will quickly scale to the next level of visibility and debug-ability across their networks.
“The right switch will offer real-time monitoring using eight smart mirror engines, which enable organisations to move from reactive to proactive mode,” clarifies Jacobsz. “These switches can mirror in real-time with nanosecond granularity in the event of drop occurrences, high latency, ACL flows and even buffer level events.”
They can also enhance debug capability using statistics, triggered watermarks and buffer status snapshots. Using histograms, IT can quickly identify microbursts and will be able to specify absolute limit values both for the data and burst size.
They offer a snapshot of the entire switch buffer’s status, which allows IT managers to quickly determine when another flow is causing starvation to a poorly performing application. Modifying these shared buffer rules can result in immediate performance boosts.
Examples of real-life use cases include low latency trading, where, for example, when executing arbitrage strategies, every millisecond lost could result in hundreds of millions of dollars in lost opportunity. Another example would lie in media streaming where poor bandwidth, delays and microburst glitches will result in a poor quality of customer experience.
“It is IT’s core challenge to ensure that the infrastructure supporting these tools delivers applications that are reliable, fast and secure,” says Jacobsz. “This means total visibility into the network and there are products available today that ensure real-time, fast, accurate over- and underlay, as well as distinct and quick debugging of fabrics to ensure the network is fully covered.”
For more information, contact Chris Coetzee at Networks Unlimited on [email protected]
© Technews Publishing (Pty) Ltd. | All Rights Reserved.
printer friendly version