The Power of Skylight Metadata: Lower TCO for Network Performance Monitoring

See how Skylight reduces TCO by 300-900%

Money matters! Especially when it comes to enterprise budgets and how they’re spent. Everyone generally agrees, there’s always far more things to do than there is budget available with which to do them. This often leads teams to select alternative ways to achieve objectives in order to cut costs and accomplish more with less, especially for network performance monitoring and application performance monitoring.

In our previous post, Experience the Power of Skylight Metadata, we discussed how Skylight metadata streamlines full network traffic packet capture data by 2000-5000%. This reduces the amount of data that needs to be sent and retained for full network and application performance visibility an history.

In this post we’ll discuss the cost benefits of Skylight’s efficient, high fidelity metadata.

Using traditional on-premise traffic visibility tools leads to unacceptable network performance monitoring costs

If that’s the case, then why would any enterprise IT ops team spend more than they need to in order to monitor networks and applications in the cloud and on-premises? Surprisingly, many enterprises are doing just that. How? By simply replicating their traditional on-premises traffic visibility strategy in the cloud. But, this creates massive cost leaks.

There are numerous limitations to this approach.

  • Cloud providers typically won’t let hardware network taps and mirror ports to be installed into cloud datacenters. That means that virtual taps need to be used to monitor east-west traffic within the cloud.
  • Virtual TAPs, which are software-defined, capture packets like hardware TAPs. They add cost and offer no intrinsic optimizations to enable all network traffic to be viewed to detect performance issues. Virtual TAPs require that all packet information be transferred to a capture appliance over the cloud network or to be sampled to reduce the sheer amount of packet traffic.
  • Full packet capture requires a high storage capacity to retain the captured packet information, either in the cloud or on-premises. If virtual tap packet filtering is used to reduce the amount of storage needed, it reduces the amount of information available to ensure precise and accurate performance monitoring.
  • Transferring packet information from the cloud to an on-premises packet capture appliance increases costs even higher. That’s because data transfer from the cloud to an external datastore is much more expensive than it is within the cloud.

The overall costs for monitoring a widespread environment using traditional tools would escalate quickly and lead to increased operating costs and significantly higher total cost of ownership (TCO). These increases could overwhelm the savings obtained by moving infrastructure to the cloud.

Different traffic visibility architectures lead to very different network performance monitoring costs

Skylight cloud network performance monitoring Competitive cloud network performance monitoring
Cloud packet broker Virtual TAP
Distributed Centralized traffic analysis
Lightweight: minimal storage requirements Heavy system and storage requirement
Rich metadata Poor metadata plus full traffic capture

But, let’s let the numbers speak for themselves.

The configuration comparisons show that a Skylight cloud monitoring configuration is between 300-900% less than comparable competitive configurations for cloud and on-premises configurations. All configurations use packet capture and stream processing to turn the packet information into high precision performance metadata in real-time.

The competitive configurations shown feature three different cloud-only configurations that vary whether every virtual private cloud (VPC) is monitored, if only regions are monitored, or if only the entire cloud configuration is monitored. The difference is that Skylight stream processing runs as an application within a VM in the cloud. The competitive configurations use hardware appliance-based stream processing that need to be connected to the VPCs using a network virtual TAP.

What’s important to note from this analysis is that, in all cases, the hosting costs for cloud configurations are comparable. They are not included for the hardware-based, on-premises configuration. As with any TCO or financial analysis, the costs will vary based upon configuration and implementation,. That’s why this analysis is proffered as a pro forma analysis. However, the configuration costs are based upon actual costs and actual requirements for each implementation.

Additional details on the costs used for this analysis:

  • The total costs don’t include the extra costs associated with virtual TAP software, if it’s required.
  • Cloud bandwidth costs also vary based upon the type of capture method use. For the competitive implementations, configurations that deploy stream processing appliances and capture for each VPC are the most economical. The configuration that sends unprocessed packet data to an external capture appliance is the most expensive.

The reason for the high cost of the on-premises data capture is because data sent from the cloud to external on-premises capture devices is tariffed at a much higher rate than for data transfer between VPCs or cloud regions.

Skylight Network Performance Monitoring Advantages

With Skylight, the amount of packet information, sent as highly efficient metadata, requires considerably less bandwidth than competitive solutions.

  • Skylight sensor software, with stream processing capabilities, also run on the same server hardware used for cloud applications. They not do not depend on more expensive dedicated stream processing hardware.
  • New sensor instances can be initiated anywhere within the cloud infrastructure within seconds. On the other hand, hardware stream processing appliances require physical installation and configuration to be added.

Overall, Skylight cloud performance monitoring costs are significantly lower than competitive implementations that use virtual TAPs or hardware-based stream processing technology. A TCO comparison for each shows the significant difference of 300-900%, dependent upon the configuration type.

That’s the end result of complete, high precision network traffic metadata that provides total network and application performance visibility in the cloud, on-premise and in hybrid environments.

But don’t just take our word for it. See what our customers are saying:

A network architect at a recruiting/HR firm with 501-1,000 employees says:

A freelance IT consultant at SPW (Service Public de Wallonie) says: