* href="https://github.com/aappleby/smhasher/blob/master/src/MurmurHash3.cpp">32-bit murmur3
   * algorithm, x86 variant</a> (little-endian variant), using the given seed value.
   *
   * <p>The exact C++ equivalent is the MurmurHash3_x86_32 function (Murmur3A).
   *
   * <p>This method is called {@code murmur3_32_fixed} because it fixes a bug in the {@code
   * HashFunction} returned by the original {@code murmur3_32} method.

   * containing one element (the given set itself).
   *
   * <p>Despite the method name, this method attempts to avoid actually copying the data when it is
   * safe to do so. The exact circumstances under which a copy will or will not be performed are
   * undocumented and subject to change.
   *
   * <p>This method is not type-safe, as it may be called on elements that are not mutually
   * comparable.

    // Break apart the domain into UTF-8 labels, i.e. foo.bar.com turns into [foo, bar, com].
    val domainLabelsUtf8Bytes = Array(domainLabels.size) { i -> domainLabels[i].encodeUtf8() }

    // Start by looking for exact matches. We start at the leftmost label. For example, foo.bar.com
    // will look like: [foo, bar, com], [bar, com], [com]. The longest matching rule wins.
    var exactMatch: String? = null

In this case, it would be better to get **one extra server** and run some processes on it so that they all have **enough RAM and CPU time**.

There's also the chance that for some reason you have a **spike** of usage of your API. Maybe it went viral, or maybe some other services or bots start using it. And you might want to have extra resources to be safe in those cases.

<img src="/img/tutorial/security/image03.png">

Não estamos verificando a validade do token ainda, mas isso já é um começo

## Recapitulando

{* ../../docs_src/custom_request_and_route/tutorial003.py hl[26] *}

Nesse exemplo, as *operações de rota* sob o `router` irão usar a classe `TimedRoute` personalizada, e terão um cabeçalho extra `X-Response-Time` na resposta com o tempo que levou para gerar a resposta:

     * Returns the duration elapsed since this clock was initialized.
     * <p>
     * The returned duration is calculated using {@link System#nanoTime()}
     * to ensure monotonic behavior. This duration represents the exact time
     * span between clock initialization and the current instant.
     *
     * @return the duration since clock initialization
     * @see #startInstant()
     * @see #instant()
     */

                    }
                }
            }).start();
        }

        startLatch.countDown();
        endLatch.await();

        // Verify all rules were added (can't check exact count due to thread-safe implementation)
        ResponseData responseData = new ResponseData();
        assertNotNull(threadSafeManager.getRule(responseData)); // At least one rule should match
    }

    /**

      //
      // We use two volatile reads. We could reduce this to one by
      // putting our fields into a holder class, but (at least on x86)
      // the extra memory consumption and indirection are more
      // expensive than the extra volatile reads.
      long nanos = expirationNanos;
      long now = System.nanoTime();
      if (nanos == 0 || now - nanos >= 0) {
        synchronized (lock) {

     *
     * <p>Description matching supports:</p>
     * <ul>
     *   <li>Wildcard matching (if starts or ends with *)</li>
     *   <li>Prefix matching (if ends with *)</li>
     *   <li>Exact phrase matching (otherwise)</li>
     * </ul>
     *
     * @param cb the condition bean to configure
     * @param dataConfigPager the pager containing search criteria
     */