What does it mean to have a local AWS region

With the latest announcement from AWS on opening of their new Region in Auckland, I thought I should share some thoughts on this news.

First of all, what is an AWS Region. AWS defines it as a physical location with cluster of datacenters. Like all of their new regions since 2016, Auckland region will also be built with 3AZs. Each one containing one DC to start with. Number of DCs in each AZ can potentially increase over the years depending on demand.

Fun fact: This is the first time Amazon has publicly released figures on establishing a new AWS region. You can find the full AWS Economic Impact Study here.

Why Now?

Organisations in NZ have been using AWS’s Sydney region, closest one to us,?for last 10 years. So, why is that AWS has made such a big decision to open a new region in Auckland. The main reason that comes to mind is to compete against Microsoft, who announced the opening of a new Azure region in Auckland last year. Without competition from AWS, Azure could have become the primary choice for SMEs and Enterprises to house their workloads, at least the ones that required to be close by. Other reason is that the barriers for shifting to public cloud have pretty much diminished over the years. One of the last, nonetheless important, reason was latency, which will be addressed with a local region. There are few others as well whcih I will talk about later in this article.

What does a local region give you?

Majority of NZ businesses have been making use of?AWS through their Sydney region and it has been working great so far. However, there are definitely some advantages to having a local region in your backyard.

Sub 5-10ms latency

Latency of 25-35ms to Sydney has been good enough for applications written with newer protocols or are latency tolerant. There are however plethora of legacy client/server applications which still require that LAN like sub 5-10ms latency. These are the ones that will truly benefit from having a local region.

We shouldn’t underestimate the importance of this low latency network. This will not only provide Enterprises the opportunity to migrate their “old” applications to a public cloud but also the SMEs who over the years have acquired fat client/server applications. Allowing them to truely exit self or co-hosted data centres.

The other hidden benefit of a low latency connections is that more throughput can be achieved over the same pipe compared to a high latency connection. You can achieve 15-20% more throughput from a 5-10ms latency connection compared to 25-30ms one. This means you can get more data through same bandwidth on a lower latency connection.

Data sovereignty

This is not a major issue anymore but still plays at the back of people’s minds. It must be in human psyche to get comfort from knowing that their data and servers are hosted in their backyard. Having a local region will help with that and minimize migration jitters.?

Shift from existing “Data Centres”

Enterprises will probably benefit the most from local public cloud presence, AWS or Azure. At last there is a real choice to completely exit fully owned or co-hosted data centers. There are however centain sytems which cannot be easily accommodated by these public clouds e.g. mainframes (yes, they are still out there) and solaris. Given the desire and benefits of exiting self managed data centres, there will be choices in the form of smaller co-hosted locations closer to these public clouds complemented with dedicated high bandwidth connections.?

This shift to public cloud can release tied up capital in managing and maintaining existing data centres. This capital can then be redeployed by organisations to modernise their business applications residing in public clouds.

?#awscloud #datacentres #cloudmigration #awsregion

?