Skip to content
7LayerX

Internet fundamentals / Hands-on lab

Ping Works, but the Application Doesn't

Learn why a successful ping does not prove a website works, then investigate DNS, routing, firewalls, ports, TLS, and the application layer in a guided simulation.

beginnerPingICMPTCP/IPDNSTLS

Lab goal

Find out why a website or service can fail even when ping gets a reply. You will follow a request through the network, compare different failure cases, and choose the next useful check.

Why ping is not enough

Ping usually sends an ICMP Echo Request and waits for an Echo Reply. A reply shows that this kind of packet reached a device that chose to answer. It does not confirm that a website is available.

A website needs more than IP reachability. The name must resolve, the route must work, the server must accept a TCP connection, TLS must complete when HTTPS is used, and the application must return a useful response. A firewall or host can also treat ping and web traffic differently.

The opposite can happen too: ping may be blocked while the website works. Treat ping as one clue, not a final health check.

The OSI layer labels in this lab are a way to organize checks. Real networks do not always fit neatly into one layer, and tools such as DNS and TLS can interact with several parts of a connection.

Before you start

  • You do not need a real network or a terminal on your computer.
  • The addresses and command output in this lab are examples.
  • The simulator does not send packets or run commands on your device.

Explore the lab

Use the controls to choose a network scenario, inspect the seven-layer view, and try the ping and trace simulations. Then select a troubleshooting step and enter example commands in the simulated shell.

Interactive lab

Maximize the lab for a focused, full-screen workspace.

7LayerX · Architecture/L1–L7 Stack Visualizer/

Trace the Signal.Understand the System.

THE CONNECTED SYSTEM · 7 LAYERSPHYSICAL TO APPLICATION
L1
Physical
Fiber, Cat6, Radio / RF, SFP+
L2
Data Link
Ethernet, VLANs, MAC, ARP
L3
NetworkPing Reaches Here
IPv4, IPv6, ICMP, Routing
L4
TransportPorts Blocked Here
TCP, UDP, Ports (80, 443)
L5
Session
NetBIOS, RPC, Sockets
L6
Presentation
TLS, SSL, JSON, Compression
L7
Application
HTTP, HTTPS, DNS, SSH, SMTP
Ping only reaches Layer 3 (Network).
Web App requires Layers 4, 6 & 7.
Diagnostic Principle

“Ping is working, but the application is not.”

Because ping primarily tests IP reachability using ICMP, while applications depend on DNS, routing, port filtering, MTU limits, TLS negotiation, and daemon execution.

Layer 7 · ApplicationPDU: Data / Messages
Typical Protocols:HTTP, HTTPS, DNS, SSH, SMTP
Role in Troubleshooting:

Where the web service and business logic run. Ping never reaches this layer.

Ping Reachability:○ Blind to Ping
App Dependency:● Critical Path

Trace the signal. Understand the system. Share the knowledge.

7LayerX · Topology Simulation·www.7layerx.com

Ping Works, but the Application Doesn't

Client 10.20.4.15 can reach 198.51.100.25 via ICMP (Layer 3), but application sockets on port 443 are denied.

Ping (ICMP) · Echo reply in this scenario
Application (TCP 443 / HTTPS) · Not Working
|
ICMP ECHO ▶TCP 443 [SYN] ▶Client10.20.4.15AccessVLAN 100Edge10.20.4.1PerimeterPolicy EngineWANPublic AS PathApplication198.51.100.25
Scenario Focus/Firewall Policy Block

You can ping 198.51.100.25 and get a reply, but the website or application doesn't open.

Network reaches server at Layer 3 (IP), but the application is blocked at a higher layer.

Root Barrier

DROPPED: ACL deny tcp any host 198.51.100.25 eq 443

Interactive Protocol Lab/Ping vs Traceroute (ICMP vs TCP)

How Ping & Traceroute Actually Work

Discover why traditional ICMP traceroute fails behind enterprise firewalls, while TCP traceroute (-T -p 443) illuminates the true application path.

Active Probe TTL:TTL = 1
Protocol:TCP SYN (Port 443)
SRCClient Host10.20.4.15H1Gateway Router10.20.4.1H2Edge Router172.16.0.1H3Perimeter Firewall198.51.100.1PERMIT 443DSTTarget Server198.51.100.25
Traceroute Execution OutputTarget: 198.51.100.25
HopIP AddressRTTResponse Mechanism
Press "Auto Run Trace" or "Next Hop" to trace packets.
Command run:traceroute -T -p 443 198.51.100.25
Wire Packet DissectionLayer 3 / 4 PDU
IPv4 Header (Layer 3)
TTL (Time to Live):1
Protocol Field:0x06 (TCP)
Source IP:10.20.4.15
Destination IP:198.51.100.25
TCP Segment (Layer 4)
Dest Port:443 (HTTPS)
Control Flags:[SYN]
Firewall Behavior:Permitted by Web Policy ACL

When a router receives an IP packet, it decrements the TTL by 1. When TTL becomes 0, the router discards the packet and returns an ICMP Time Exceeded message to the sender.

Deep Dive: How ICMP Ping and Traceroute Function Under the Hood

1. How Ping Works

Ping operates purely on ICMP (Protocol 1) encapsulated in raw IP packets. It transmits an ICMP Echo Request (Type 8) with a timestamp and sequence number.

When the target host's kernel receives it, it replies with an ICMP Echo Reply (Type 0). Because ping contains no TCP or UDP port headers, it cannot verify whether port 443 is open or if Nginx is running!

2. The TTL Exploitation Trick

Traceroute works by intentionally crafting IP packets that expire. It sends probe 1 with TTL=1. The first router decrements TTL to 0, drops the packet, and responds with ICMP Type 11 Code 0 (Time to Live Exceeded).

Next, it sends probe 2 with TTL=2, reaching the second router before expiring. By incrementing TTL one by one, traceroute maps every router hop along the path.

3. Why TCP Traceroute is Critical

Standard Windows tracert uses ICMP, and Linux uses UDP. Both frequently hit firewalls that silently drop the probes, resulting in endless * * * lines.

TCP Traceroute (-T -p 443) sends real TCP SYN packets targeting HTTPS port 443. Firewalls permit it through because web traffic is authorized, revealing the true path right to the application daemon!

MethodTransport ProtocolProbe HeaderFirewall TraversalBest Used For
PingICMP (Proto 1)Type 8 (Echo)Permitted on basic networks; blocked on hardened DMZBasic Layer 3 reachability & packet loss check
ICMP TracerouteICMP (Proto 1)Type 8 (TTL 1..N)Frequently blocked by stateful firewalls (* * *)Quick internal LAN routing discovery (Windows default)
TCP TracerouteTCP (Proto 6)TCP SYN :443 / :80May reveal hops when TCP probes are allowedDiagnosing web application connectivity failures
UDP TracerouteUDP (Proto 17)High Port 33434+May be filtered by firewalls or routersClassic Unix / Cisco core router path discovery

Systematic Troubleshooting Protocol

Diagnostic verification across layers 1 through 7

1

Check DNS Resolution

Layer

Verify the domain name resolves to the correct IP address.

nslookup example.com
2

Check Port Connectivity

Layer

Ping only checks ICMP. Applications require specific TCP or UDP ports.

nc -zv 198.51.100.25 443
3

Verify Firewall Rules

Layer

Ensure firewalls allow the required application ports.

show access-lists | include 198.51.100.25
4

Check Routing (Both Directions)

Layer

Ensure the return path is valid and there is no asymmetric routing.

traceroute -T -p 443 198.51.100.25
5

Check MTU / Fragmentation

Layer

Large packets may be dropped, affecting application traffic.

ping -M do -s 1472 198.51.100.25
6

Verify TLS/SSL (for HTTPS)

Layer

Check certificate issues, TLS handshake failures, or unsupported versions.

openssl s_client -connect 198.51.100.25:443 -servername example.com
7

Check Proxy or Application Settings

Layer

Ensure correct proxy configuration and security filtering policies.

curl -x http://proxy.corp:8080 -I https://example.com
8

Verify The Application Service

Layer

Ensure the service is running and listening on the server IP & port.

sudo systemctl status nginx
1Layer 7 (Application)

Check DNS Resolution

If users access the application using a hostname, verify that it resolves to 198.51.100.25. If the hostname fails to resolve, investigate DNS first. A direct-IP test is useful, but HTTPS virtual hosting and certificates can make it behave differently from a hostname request.

Diagnostic Signature:

Check if the returned IP matches 198.51.100.25. Look out for NXDOMAIN, DNS timeouts, or stale CDN anycast IPs.

Verified Commands:Click to copy
Cross-platformnslookup example.com
Linux / macOSdig +short example.com
PowerShellResolve-DnsName example.com
Execution Sample:
$ nslookup example.com
Server:         1.1.1.1
Address:        1.1.1.1#53

Non-authoritative answer:
Name:   example.com
Address: 198.51.100.25

Simulated diagnostics shell

Run example commands against the lab scenario. No network probes are sent.

|
noc@7layerx:~$ (10.20.4.15 → 198.51.100.25)
Probes:
7LayerX Shell · Target: 198.51.100.25 · Source: 10.20.4.15
Verify reachability layer-by-layer. Type 'help' or run preset probes.
[02:20:01]$ping 198.51.100.25 -c 4
PING 198.51.100.25 (198.51.100.25) 56(84) bytes of data from 10.20.4.15.
64 bytes from 198.51.100.25: icmp_seq=1 ttl=58 time=2.14 ms
64 bytes from 198.51.100.25: icmp_seq=2 ttl=58 time=2.08 ms
64 bytes from 198.51.100.25: icmp_seq=3 ttl=58 time=2.11 ms
64 bytes from 198.51.100.25: icmp_seq=4 ttl=58 time=2.05 ms

--- 198.51.100.25 ping statistics ---
4 packets transmitted, 4 received, 0% packet loss, time 3004ms
rtt min/avg/max/mdev = 2.050/2.095/2.140/0.034 ms

[INFO] This example received an ICMP Echo Reply from 198.51.100.25. Other services still need separate checks.
[02:20:06]$nc -zv -w 3 198.51.100.25 443
nc: connect to 198.51.100.25 port 443 (tcp) timed out: Operation in progress
[INFO] The TCP connection timed out in this example. Filtering, a path issue, or an unavailable service could cause this; check further.
$

Why Ping Works but the Application Fails

Comparative failure vectors across the OSI stack

ReasonLayerTechnical Root Cause & Action
Different ProtocolsLayer 3 vs Layer 4

Ping uses ICMP (Protocol 1), whereas applications use TCP (6) or UDP (17).

Action: Test port 443 with netcat or telnet rather than relying on ping.

Firewall Blocking PortsLayer 4 Security

Some firewalls allow ICMP echo while filtering one or more application ports.

Action: Verify security policy ACLs on routers and firewalls.

DNS IssuesLayer 7 Resolution

Domain resolves to wrong IP, has expired TTL, or fails internal resolution.

Action: Run nslookup or dig to confirm hostname-to-IP accuracy.

Application Service DownLayer 7 Host Service

One possible case: the server responds to ICMP, but its web service is stopped or not listening.

Action: Check systemctl status and listening sockets (ss -tulpn).

TLS/SSL ProblemsLayer 6 Presentation

Certificate errors, expired cert, handshake cipher mismatch, or SNI misconfiguration.

Action: Test with openssl s_client -connect host:443.

Routing AsymmetryLayer 3 Routing

The return path differs, which can cause problems when stateful devices expect to see both directions.

Action: Compare traces from both ends and check stateful firewall logs.

MTU / FragmentationLayer 2/3 MTU

Small ICMP packets pass, but large application packets exceed MTU and are dropped.

Action: Test with ping -M do -s 1472 (DF bit set).

Proxy or Security PoliciesLayer 7 Proxy/WAF

Proxy settings, URL filtering, or security policies blocking application payloads.

Action: Check browser proxy config and proxy server logs.

Application MisconfigurationLayer 7 App Config

Incorrect virtual host binding, listening on 127.0.0.1 instead of 0.0.0.0, or 500 error.

Action: Inspect web server virtual host configs and application logs.

Core Engineering Takeaways

This ping received an ICMP Echo Reply: It does not test DNS, a service port, TLS, or the application.

An application connection may also depend on: DNS + Bidirectional Routing + TCP/UDP Ports + Firewall ACLs + Path MTU + TLS/SSL + Application Daemons.

Troubleshoot layer-by-layer: Ascend methodically from Physical/Link to Network, Transport, and Application layers.

Key Takeaway: A successful ping does NOT mean the application will work.

Common Application Ports

ServiceProtocol / Port
HTTPTCP 80
HTTPSTCP 443
DNSUDP/TCP 53
SMTPTCP 25
POP3TCP 110
IMAPTCP 143
RDPTCP 3389
SSHTCP 22

Unlike ICMP, firewall state engines inspect and enforce access control on these exact Layer 4 port tuples.

This is an interactive simulation. It does not connect to your device, network, or the example destination.

How to read the results

Compare what ping reports with what the application needs. If ping succeeds but a TCP connection to the service fails, check the service port, route, firewall policy, and server listener. If the connection works but the page still fails, check DNS, TLS, proxy settings, and the application response.

Do not assume a timeout proves that a firewall is the cause. A timeout can have several causes. Use a second check and compare results from more than one point when possible.

Choose a useful next check

What you seeA useful next check
The hostname does not resolveCheck the configured DNS server and the DNS response.
Ping replies, but a TCP connection times outCheck the route, filtering, and whether the service is listening. The timeout alone does not identify the cause.
TCP connects, but HTTPS failsInspect the TLS handshake, certificate, and hostname.
HTTPS connects, but the page is wrong or unavailableCheck the HTTP response, proxy, and application logs.

These are starting points, not proof of a root cause. Compare the results with the expected behavior for your service.

Key takeaway

Ping tests one type of network reachability. It does not test every step needed by an application. Check the name, path, service port, secure connection, and application response to narrow down where the failure begins.

Continue learning

Read the Internet Fundamentals topic or review The Seven Layers of Modern Internet Communication for a broader view of how a web request moves through the stack.

Pass it along

Found this useful? Share it.

Choose a platform or copy the direct link.

LinkedInXFacebook

Continue reading

Reader signal

Was this useful?

Your feedback helps shape the next field note and lab.

… views

Views count when this page opens. You can change or remove your vote during this tab session.