Latency between ElasticSearch and AWS

I have an Amazon EC2 instance used to make requests to an Elastic Cloud instance hosted on AWS. Both are in the same region on AWS (N. Virginia). However, I am seeing incredibly high ping between the two servers.

Using the ElasticSearch PHP client, timing client->ping() measures the time at about 200 milliseconds.

I am trying to implement search-as-you-type. When I time the queries in Kibana, they take only a few milliseconds to complete. However, the search-as-you-type is unusable due to the latency between ES and EC2 servers.

How can I diagnose/resolve this issue? I would expect a ping of around a tenth of what I am getting due to the proximity of the servers.

Are you including the time to initialise the client and/or set up the connection in that latency?