Archives

Posts Tagged ‘Network Performance Management’

0

3 Key Differences Between NetFlow and Deep Packet Inspection (DPI) Packet Capture Monitoring

May 5, 2023 Rafi Sabel / / / / / / / / / / / / /

The increasing density, complexity and expanse of modern networking environments have fueled the ongoing debate around which network analysis and monitoring tools serve the needs of the modern engineer best – placing Packet Capture and NetFlow Analysis to perform NDR at center-stage of the conversation.

Granted, when performing analysis of unencrypted traffic both can be extremely valuable tools in ongoing efforts to maintain and optimize complex environments, but as an engineer, I tend to focus on solutions that give me the insights I need without too much cost on my resources, while complementing my team’s ability to maintain and optimize the environments we support.

So with this in mind, let’s take a look at how NetFlow, in the context of the highly-dense networks we find today, delivers three key requirements network teams rely on for reliable end-to-end performance monitoring of their environments.

A NetFlow deployment won’t drain your resources

Packet Capture, also known as Deep Packet Inspection (DPI), once rich in network metrics has finally failed due to encryption and a segment based approach making it expensive to deploy and maintain. It requires a requires sniffing devices and agents throughout the network, which invariably require a huge of maintenance during their lifespan.

In addition, the amount of space required to store and analyze packet data makes it an inefficient an inelegant method of monitoring or forensic analysis. Combine this with the levels of complexity networks can reach today, and overall cost and maintenance associated with DPI can quickly become unfeasible. In the case of NetFlow, its wide vendor support across virtually the entire networking landscape makes almost every switch, router, vmware, GCP cloud, Azure Cloud, AWS cloud, vmWare velocloud or firewall a NetFlow / IPFIX / sflow / ixflow “ready” device. Devices’ built-in readiness to capture and export data-rich metrics makes it easy for engineers to deploy and utilize . Also, thanks to its popularity, CySight’s NetFlow analyzer provides varying feature-sets with enriched vendor specific flow fields are available for security operations center (SOC) network operations center (NOC) teams to gain full advantage of data-rich packet flows.

Striking the balance between detail and context

Considering how network-dependent and widespread applications have become in recent years, NetFlow, IPFIX, sFlow and ixFlow’s  ability to provide WAN-wide metrics in near real-time makes it a  suitable troubleshooting companion for engineers. Add to this enriched context enables a very complete qualification of impact from standard traffic analysis perspective as well as End point Threat views and Machine Learning and AI-Diagnostics.

Latest Flow methods extend the wealth of information as it collects via a template-based collection scheme, it strikes the balance between detail and high-level insight without placing too much demand on networking hardware – which is something that can’t be said for Deep Packet Inspection. Netflow’s constant evolution alongside the networking landscape is seeing it used as a complement to solutions such as Cisco’s NBAR , Packet Brokers such as KeySight, Ixia, Gigamon, nProbe, NetQuest, Niagra Networks, CGS Tower Networks, nProbe and other Packet Broker solutions have recognized that all they need to export flexible enriched flow fields to reveal details at the packet level.

NetFlow places your environment in greater context

Context is a chief area where Granular NetFlow beats out Packet Capture since it allows engineers to quickly locate root causes relating to Cyber Security, Threat Hunting, Root Cause and Performance by providing a more situational view of the environment, its data-flows, bottleneck-prone segments, application behavior, device sessions and so on.

One could argue that Deep Packet Inspection (DPI) is able to provide much of this information too, but as networks today are over 98% encrypted even using certificates won’t give engineers the broader context around the information it presents, thus hamstringing IT teams from detecting anomalies that could be subscribed to a number of factors such as cyber threats, untimely system-wide application or operating system updates or a cross-link backup application pulling loads of data across the WAN during operational hours.

So does NetFlow make Deep Packet Inspection obsolete?

Both Deep Packet Inspection (DPI) and legacy Netflow Analyzers cannot scale in retention so when comparing those two genres of solutions the only win a low end netflow analyzer solution will have against a DPI solution is that DPI is segment based so flow solution is inherently better as its agentless.

However, using NetFlow to identify an attack profile or illicit traffic can only be attained when flow retention is deep (granular) . However, NetFlow strikes that perfect balance between detail and context and gives SOC’s and NOCs intelligent insights that reveals broader factors that can influence your network’s ability to perform.

Gartner’s assertion that a balance of 80% NetFlow monitoring  coupled with 20% Packet Capture as the perfect combination of performance monitoring is false due to encryption’s rise but it is correct to attest to NetFlow’s growing prominence as the monitoring tool of choice and as it and its various iterations such sFlow, IPFIX, ixFlow, Flow logs and  others flow protocols continue to expand the breadth of context it provides network engineers, that margin is set to increase in its favor as time.

8 Keys to Understanding NetFlow for Network Security, Performance & Overall IT Health

0

5 Benefits of NetFlow Performance Monitoring

June 28, 2021 Rafi Sabel / / / / /

NetFlow can provide the visibility that reduces the risks to business continuity associated with poor performance, downtime and security threats. As organizations continue to rely more and more on computing power to run their business, transparency of business applications across the network and interactions with external systems and users is critical and NetFlow monitoring simplifies detection, diagnosis, and resolution of network issues.

Traditional SNMP tools are too limited in their collection capability and the data extracted is very coarse and does not provide sufficient visibility. Packet capture tools can extract high granularity from network traffic but are dedicated to a port mirror or a per segment capture and are built for specialized short term captures rather than long historical analyses as well as packet based Predictive AI Baselining analytics tools traditionally do not provide the ability to perform forensic analysis of various sub-sections or aggregated views and therefore also suffer from a lack of visibility and context.

By analyzing NetFlow data, a network engineer can discover the root cause of congestion and find out the users, applications and protocols that are causing interfaces to exceed utilization. They can determine the Type or Class of Service (QoS) and group with application mapping such as Network Based Application Recognition (NBAR) to identify the type of traffic such as when peer-to-peer traffic tries to hide itself by using web services and so on.

NetFlow allows granular and accurate traffic measurements as well as high-level aggregated traffic collection that can assist in identifying excessive bandwidth utilization or unexpected application traffic. NetFlow enables networks to perform IP traffic flow analyses without deploying external probes, making traffic Predictive AI Baselining analytics cheap to deploy even on large network environments.

Understanding the benefits of a properly managed network might be a cost that hasn’t been considered by business managers.  So what are the benefits of NetFlow performance monitoring?

Avoiding Downtime

Service outages can cause simple frustrations such as issues when saving opened files to a suddenly inaccessible server or cause delays to business process can impact revenue. Downtime that affects customers can result in negative customer experiences and lost revenue.

Network Speed

Slow network links cause severe user frustration that impacts productivity and causes delays.

Scalability

As the business grows, the network needs to grow with it to support more computing processes, employees and clients. As performance degrades, it can be easy to set thresholds that show when, where and why the network is exceeding it’s capability. By having historical performance of the network, it is very easy to see when or where it is being stretched too thin or overwhelmed and to easily substantiate upgrades.

Security

Security Predictive AI Baselining analytics is arguably the most important aspect of network management, even though it might not be thought of as a performance aspect. By monitoring NetFlow performance, it’s easy to see where the most resources are being used. Many security attacks drain resources, so if there are resource spikes in unusual areas it can point to a security flaw. With advanced NetFlow diagnostics software, these issues can be not only monitored, but also recorded and corrected.

Peering

NetFlow Predictive AI Baselining analytics allows organizations to embrace progressive IT trends to analyze peering paths, or virtualized paths such as Multiprotocol Label Switching (MPLS), Virtual Local Area Networks (VLAN), make use of IPv6, Next Hops, BGP, MAC Addresses or link internal IP’s to Network Address Translated (NAT) IP addresses.

MPLS creates new challenges when it comes to network performance monitoring and security with MPLS as communication can occur directly without passing through an Intrusion Detection System (IDS). Organizations have two options to address the issue: Implement a sensor or probe at each site, which is costly and fraught with management hassles; or turn on NetFlow at each MPLS site and send MPLS flows to a NetFlow collector that becomes a very cost-effective option because it can make use of existing network infrastructure.

With NetFlow, network performance issues are easily resolved because engineers are given in-depth visibility that enables troubleshooting – all the way down the network to a specific user. The solution starts by presenting the interfaces being used.

An IT engineer can click on the interfaces to determine the root cause of a usage spike, as well as the actual conversation and host pairs that are dominating bandwidth and the associated user information. Then, all it takes is a quick phone call to the end-user, instructing them to shut down resource-hogging activities, as they’re impairing network performance.

IT departments are tasked with responding to end-user complaints, which often relate to network sluggishness. However, many times an end-user’s experience has more to do with an application or server failing to respond within the expected time period that can point to a Latency, Quality of Service (QoS) or server issue. Traffic Latency can be analyzed if supported in the NetFlow version. Statistics such as Round Trip Time (RTT) and Server Response Time (SRT) can be diagnosed to identify whether a network problem is due to an application or service experiencing slow response times.

Baselines of RTT and SRT can be learned to determine whether response time is consistent or is spiking over a time period. By observing the Latency over time alarms can be generated when Latency increases above the normal threshold.

TCP Congestion flags can also serve as an early warning system where Latency is unavailable.

NetFlow application performance monitoring provides the context around an event or anomaly within a host to quickly identify the root cause of issues and improve Mean Time To Know (MTTK) and Mean Time To Repair (MTTR).

Performance Monitoring & Security Forensics: The 1-2 Punch for Network and IT Infrastructure Visibility

0

Two Ways Networks Are Transformed By NetFlow

April 21, 2021 Rafi Sabel / / / / / / / / / / / / / / / / / / /

According an article in techtarget.com “Your routers and switches can yield a mother lode of information about your network–if you know where to dig.”  The article goes on to say that excavating and searching through endless traffic data and logs manufactured by your network system is a lot like mining for gold, and punching random holes to look for a few nuggets of information isn’t very efficient. Your search will be much more fruitful if you know where to look and what it will look like. Fortunately, the data generated by a NetFlow traffic reporting protocol yields specific information and you can easily sort, view and analyze the information into what you want to use or need.In contemporary networks, there is a need to collect and retain a good set of traffic records for several different purposes. These include the ability to monitor traffic for network planning, security and analysis as well as track traffic usage for billing purposes. Every business experiences network problems. The goal is to transform these “bad behaving” networks by investigating the data that is being generated by the routers, switches and other hardware that make up the system.

  • Trace and repair network misconfigurations

Problems with networks can run the gamut from mismatched applications and hardware to wireless access points opened to accommodate BYOD users and other business uses. While there is always talk about software flaws and news about the latest internet threat, those things often distract IT pros from the real, every-day threat of unstable networks that have been configured to accommodate legacy hardware and a multitude of software applications.

The increasing complexity of the Internet itself, with the interconnection of lots of different devices and device types adds to the challenge of operating a computer network. Even though developing protocols to respond to unpredicted failures and misconfigurations is a workable solution, these out-of-date configurations can still cause frequent problems and denial of service (DOS). With many modern network devices monitoring functions and gathering data, retrieving and utilizing the NetFlow information makes tracing and repairing the problem of misconfigurations possible, easier and efficient.

  • Detect security breaches

There are many uses for NetFlow but one of the most important is the benefit of network security. This quote from an article by Wagner and Bernhard, describing worm and anomaly detection in fast IP networks, bears out the security problems facing governments, businesses, and internet users today.

“Large-scale network events, such as outbreaks of a fast Internet worm are difficult to detect in real-time from observed traffic, even more so if the observed network carries a large amount of traffic. Even during worm outbreaks, actual attack traffic is only a small fraction of the overall traffic. Its precise characteristics are usually unknown beforehand, making direct matching hard. It is desirable to derive parameters that describe large amounts of traffic data in such a way that details are hidden, but the type of changes associated with worm outbreaks and other network events are visible.”

NetFlow provides a 24/7 account of all network activity. There is an “unblinking” eye observing anything and everything that happens within the network boundaries. All the data needed to identify and enact a clean-up is recorded in the flow and this is invaluable to a security pro trying to reduce the impact of a breach in the network. NetFlow provides a visible, “what’s happening right now” view that other systems cannot provide. Most security systems alert after something has been detected, while NetFlow is constantly gathering information even when things seem to be just fine. In addition, NetFlow-based analysis relies on traffic behavior and algorithms which provides rapid detection of breaches that other technologies often miss

8 Keys to Understanding NetFlow for Network Security, Performance & Overall IT Health

0

Benefits of a NetFlow Performance Deployment in Complex Environments

March 13, 2021 Rafi Sabel / / / / / / / / / / / / /

Since no two environments are identical and no network remains stagnant in Network Monitoring today, the only thing we can expect is the unexpected!

The network has become a living dynamic and complex environment that requires a flexible approach to monitor and analyze. Network and Security teams are under pressure to go beyond simple monitoring techniques to quickly identify the root causes of issues, de-risk hidden threats and to monitor network-connected things.

A solution’s flexibility refers to not only its interface but also the overall design.

From a user interface perspective, flexibility refers to the ability to perform analysis on any combination of data fields with multiple options to view, sort, cut and count the analysis.

From a deployment perspective, flexibility means options for deployment on Linux or Windows environments and the ability to digest all traffic or scale collection with tuning techniques that don’t fully obfuscate the data.

Acquiring flexible tools are a superb investment as they enrich and facilitate local knowledge retention. They enable multiple network centric teams to benefit from a shared toolset and the business begins to leverage the power of big data Predictive AI Baselining analytics that, over time, grows and extends beyond the tool’s original requirements as new information becomes visible.

What makes a Network Management System (NMS) truly scalable is its ability to analyze all the far reaches of the enterprise using a single interface with all layers of complexity to the data abstracted.

NetFlow, sFlow, IPFIX and their variants are all about abstracting routers, switches, firewalls or taps from multiple vendors into a single searchable network intelligence.

It is critical to ensure that abstraction layers are independently scalable to enable efficient collection and be sufficiently flexible to enable multiple deployment architectures to provide low-impact, cost-effective solutions that are simple to deploy and manage.

To simplify deployment and management it has to work out the box and be self-configuring and self-healing. Many flow monitoring systems require a lot of time to configure or maintain making them expensive to deploy and hard to use.

A flow-based NMS needs to meet various alerting, Predictive AI Baselining analytics, and architectural deployment demands. It needs to adapt to rapid change, pressure on enterprise infrastructure and possess the agility needed to adapt at short notice.

Agility in provisioning services, rectifying issues, customizing and delivering alerts and reports and facilitating template creation, early threat detection and effective risk mitigation, all assist in propelling the business forward and are the hallmarks of a flexible network management methodology.

Here are some examples that require a flexible approach to network monitoring:

  • DDoS attack behavior changes randomly
  • Analyze Interface usage by Device by Datacenter by Region
  • A new unknown social networking application suddenly becomes popular
  • Compliance drives need to discover Insider threats and data leakages occurring under the radar
  • Companies grow and move offices and functions
  • Laws change requiring data retention suitable for legal compliance
  • New processes create new unplanned pressures
  • New applications cause unexpected data surges
  • A vetted application creates unanticipated denials of service
  • Systems and services become infected with new kinds of malicious agents
  • Virtualization demands abruptly increase
  • Services and resources require a bit tax or 95th percentile billing model
  • Analyzing flexible NetFlow fields supported by different device vendors such as IPv6, MPLS, MAC, BGP, VPN, NAT paths, DNS, URL, Latency etc.
  • Internet of Things (IoT) become part of the network ecosystem and require ongoing visibility to manage

Performance Monitoring & Security Forensics: The 1-2 Punch for Network and IT Infrastructure Visibility