TLS Origination for Egress Traffic
This task uses the new v1alpha3 traffic management API. The old API has been deprecated and will be removed in the next Istio release. If you need to use the old version, follow the docs here.
The Control Egress Traffic task demonstrates how external (outside the Kubernetes cluster) HTTP and HTTPS services can be accessed from applications inside the mesh. A quick reminder: by default, Istio-enabled applications are unable to access URLs outside the cluster. To enable such access, a ServiceEntry for the external service must be defined, or, alternatively, direct access to external services must be configured.
This task describes how to configure Istio to perform TLS origination for egress traffic.
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, such a requirement can be achieved just by configuration, without changing the code of the application.
In this task we show how to configure Istio to open HTTPS connections to external services in cases the original traffic was HTTP. The application will send unencrypted HTTP requests as previously and Istio will encrypt the requests for the application.
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, do
$ kubectl apply -f @samples/sleep/sleep.yaml@
otherwise, you have to manually inject the sidecar before deploying the
sleep
application:$ kubectl apply -f <(istioctl kube-inject -f @samples/sleep/sleep.yaml@)
Note that any pod that you can
exec
andcurl
from would do. -
Create a shell variable to hold the name of the source pod for sending requests to external services. If we used the sleep sample, we run:
$ export SOURCE_POD=$(kubectl get pod -l app=sleep -o jsonpath={.items..metadata.name})
Configuring HTTP and HTTPS external services
First, we configure access to cnn.com in the same way as in the Control Egress Traffic task.
Note that we use a wildcard *
in our hosts
definition: *.cnn.com
. Using the wildcard we allow access to www.cnn.com as well as to edition.cnn.com.
-
Create a
ServiceEntry
to allow access to an external HTTP and HTTPS services:cat <<EOF | istioctl create -f - apiVersion: networking.istio.io/v1alpha3 kind: ServiceEntry metadata: name: cnn spec: hosts: - "*.cnn.com" ports: - number: 80 name: http-port protocol: HTTP - number: 443 name: https-port protocol: HTTPS 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 (unimportant details replaced by ellipsis):
Note the -L
flag of curl
. It instructs curl
to follow redirects. In this case,
the server responded with a redirect response (301 Moved Permanently) to an HTTP request to http://edition.cnn.com/politics. The redirect response instructs the client to send an additional request, this time by HTTPS to https://edition.cnn.com/politics. For the second request, the server responds with the requested content and 200 OK status code.
While for the user of curl
this redirection happens transparently, there are two issues here. The first issue is the redundant first 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 our application and cnn.com, the attacker would know which specific topics and articles of cnn.com our application fetched. Due to the privacy reasons we may want to prevent such disclosure from the attacker.
In the next section we configure Istio to perform TLS origination to resolve the two issues above. Let's clean our configuration before proceeding to the next section:
$ istioctl delete serviceentry cnn
TLS origination for Egress traffic
-
Define a
ServiceEntry
to allow traffic to edition.cnn.com, aVirtualService
to perform request port rewriting, and aDestinationRule
for TLS origination.Unlike the ServiceEntry in the previous section, here we use HTTP for the protocol on port 433, since clients will send HTTP requests and Istio will perform TLS origination for them. Also, the resolution must be set to DNS to correctly configure Envoy in this case.
Finally, note that the VirtualService uses a specific host edition.cnn.com (no wildcard) because the Envoy proxy needs to know exactly which host to access using HTTPS.
cat <<EOF | istioctl create -f - apiVersion: networking.istio.io/v1alpha3 kind: ServiceEntry metadata: name: cnn 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: rewrite-port-for-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: originate-tls-for-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
-
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 we receive 200 OK as the first and the only response. Istio performed TLS origination for curl
so the original HTTP request was forwarded to cnn.com as HTTPS. The server of cnn.com returned the content directly, without the need for redirection. We spared the double round trip between the client and the server, and the request left the mesh encrypted, without disclosing the fact that our application fetched the politics section of cnn.com.
Note that we used the same command as in the previous section. For applications that access external services programmatically, the code will not be changed. We get the benefits of TLS origination by configuring Istio, without changing a line of code, transparently for the application.
Additional security considerations
Note that the traffic between the application pod and the sidecar proxy on the local host is still unencrypted. It means that if the attackers would be able to penetrate the node of our application, they would still be able to see the unencrypted communication on the local network of the node. In some environments a strict security requirement may exist that would state that all the traffic must be encrypted, even on the local network of the nodes. With such a strict requirement the applications should use HTTPS (TLS) only, the TLS origination described in this task will not be sufficient.
Also note that even for HTTPS originated by the application, the attackers could know that the requests to 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, it does not prevent the attackers from learning that cnn.com is accessed.
Cleanup
-
Remove the Istio configuration items we created:
$ istioctl delete serviceentry cnn $ istioctl delete virtualservice rewrite-port-for-edition-cnn-com $ istioctl delete destinationrule originate-tls-for-edition-cnn-com
-
Shutdown the sleep service:
$ kubectl delete -f @samples/sleep/sleep.yaml@
What's next
- Learn more about service entries, virtual services and destination rules.