TLS Origination for Egress Traffic

6 minute read

The Control Egress Traffic task demonstrates how external, i.e., outside of the service mesh, HTTP and HTTPS services can be accessed from applications inside the mesh. As described in that task, a ServiceEntry is used to configure Istio to access external services in a controlled way. This example shows how to configure Istio to perform TLS origination for traffic to an external service. Istio will open HTTPS connections to the external service while the original traffic is HTTP.

Use case

Consider a legacy application that performs HTTP calls to external sites. Suppose the organization that operates the application receives a new requirement which states that all the external traffic must be encrypted. With Istio, this requirement can be achieved just by configuration, without changing any code in the application. The application can send unencrypted HTTP requests and Istio will then encrypt them for the application.

Another benefit of sending unencrypted HTTP requests from the source, and letting Istio perform the TLS upgrade, is that Istio can produce better telemetry and provide more routing control for requests that are not encrypted.

Before you begin

Setup Istio by following the instructions in the Installation guide.
Start the sleep sample which will be used as a test source for external calls.
If you have enabled automatic sidecar injection, deploy the sleep application:
Zip
```
$ kubectl apply -f @samples/sleep/sleep.yaml@
```
Otherwise, you have to manually inject the sidecar before deploying the sleep application:
Zip
```
$ kubectl apply -f <(istioctl kube-inject -f @samples/sleep/sleep.yaml@)
```
Note that any pod that you can exec and curl from will do for the procedures below.
Create a shell variable to hold the name of the source pod for sending requests to external services. If you used the sleep sample, run:
```
$ export SOURCE_POD=$(kubectl get pod -l app=sleep -o jsonpath={.items..metadata.name})
```

Configuring access to an external service

First start by configuring access to an external service, edition.cnn.com, using the same technique shown in the Control Egress Traffic task. This time, however, use a single ServiceEntry to enable both HTTP and HTTPS access to the service.

Create a ServiceEntry and VirtualService to enable access to edition.cnn.com:

$ kubectl apply -f - <<EOF
apiVersion: networking.istio.io/v1alpha3
kind: ServiceEntry
metadata:
  name: edition-cnn-com
spec:
  hosts:
  - edition.cnn.com
  ports:
  - number: 80
    name: http-port
    protocol: HTTP
  - number: 443
    name: https-port
    protocol: HTTPS
  resolution: DNS
---
apiVersion: networking.istio.io/v1alpha3
kind: VirtualService
metadata:
  name: edition-cnn-com
spec:
  hosts:
  - edition.cnn.com
  tls:
  - match:
    - port: 443
      sni_hosts:
      - edition.cnn.com
    route:
    - destination:
        host: edition.cnn.com
        port:
          number: 443
      weight: 100
EOF

Make a request to the external HTTP service:

$ kubectl exec -it $SOURCE_POD -c sleep -- curl -sL -o /dev/null -D - http://edition.cnn.com/politics
HTTP/1.1 301 Moved Permanently
...
location: https://edition.cnn.com/politics
...

HTTP/1.1 200 OK
Content-Type: text/html; charset=utf-8
...
Content-Length: 151654
...

The output should be similar to the above (some details replaced by ellipsis).

Notice the -L flag of curl which instructs curl to follow redirects. In this case, the server returned a redirect response (301 Moved Permanently) for the HTTP request to http://edition.cnn.com/politics. The redirect response instructs the client to send an additional request, this time using HTTPS, to https://edition.cnn.com/politics. For the second request, the server returned the requested content and a 200 OK status code.

Although the curl command handled the redirection transparently, there are two issues here. The first issue is the redundant request, which doubles the latency of fetching the content of http://edition.cnn.com/politics. The second issue is that the path of the URL, politics in this case, is sent in clear text. If there is an attacker who sniffs the communication between your application and edition.cnn.com, the attacker would know which specific topics of edition.cnn.com the application fetched. For privacy reasons, you might want to prevent such disclosure.

Both of these issues can be resolved by configuring Istio to perform TLS origination.

TLS origination for egress traffic

Redefine your ServiceEntry and VirtualService from the previous section to rewrite the HTTP request port and add a DestinationRule to perform TLS origination.

$ kubectl apply -f - <<EOF
apiVersion: networking.istio.io/v1alpha3
kind: ServiceEntry
metadata:
  name: edition-cnn-com
spec:
  hosts:
  - edition.cnn.com
  ports:
  - number: 80
    name: http-port
    protocol: HTTP
  - number: 443
    name: http-port-for-tls-origination
    protocol: HTTP
  resolution: DNS
---
apiVersion: networking.istio.io/v1alpha3
kind: VirtualService
metadata:
  name: edition-cnn-com
spec:
  hosts:
  - edition.cnn.com
  http:
  - match:
    - port: 80
    route:
    - destination:
        host: edition.cnn.com
        port:
          number: 443
---
apiVersion: networking.istio.io/v1alpha3
kind: DestinationRule
metadata:
  name: edition-cnn-com
spec:
  host: edition.cnn.com
  trafficPolicy:
    loadBalancer:
      simple: ROUND_ROBIN
    portLevelSettings:
    - port:
        number: 443
      tls:
        mode: SIMPLE # initiates HTTPS when accessing edition.cnn.com
EOF

As you can see, the VirtualService redirects HTTP requests on port 80 to port 443 where the corresponding DestinationRule then performs the TLS origination. Notice that unlike the ServiceEntry in the previous section, this time the protocol on port 443 is HTTP, instead of HTTPS. This is because clients will only send HTTP requests and Istio will upgrade the connection to HTTPS.

Send an HTTP request to http://edition.cnn.com/politics, as in the previous section:
```
$ kubectl exec -it $SOURCE_POD -c sleep -- curl -sL -o /dev/null -D - http://edition.cnn.com/politics
HTTP/1.1 200 OK
Content-Type: text/html; charset=utf-8
...
Content-Length: 151654
...
```
This time you receive 200 OK as the first and the only response. Istio performed TLS origination for curl so the original HTTP request was forwarded to edition.cnn.com as HTTPS. The server returned the content directly, without the need for redirection. You eliminated the double round trip between the client and the server, and the request left the mesh encrypted, without disclosing the fact that your application fetched the politics section of edition.cnn.com.
Note that you used the same command as in the previous section. For applications that access external services programmatically, the code does not need to be changed. You get the benefits of TLS origination by configuring Istio, without changing a line of code.

Additional security considerations

Because the traffic between the application pod and the sidecar proxy on the local host is still unencrypted, an attacker that is able to penetrate the node of your application would still be able to see the unencrypted communication on the local network of the node. In some environments a strict security requirement might state that all the traffic must be encrypted, even on the local network of the nodes. With such a strict requirement, applications should use HTTPS (TLS) only. The TLS origination described in this example would not be sufficient.

Also note that even with HTTPS originated by the application, an attacker could know that requests to edition.cnn.com are being sent by inspecting Server Name Indication (SNI). The SNI field is sent unencrypted during the TLS handshake. Using HTTPS prevents the attackers from knowing specific topics and articles but does not prevent an attackers from learning that edition.cnn.com is accessed.

Cleanup

Remove the Istio configuration items you created:

$ kubectl delete serviceentry edition-cnn-com
$ kubectl delete virtualservice edition-cnn-com
$ kubectl delete destinationrule edition-cnn-com

Shutdown the sleep service:

$ kubectl delete -f @samples/sleep/sleep.yaml@