Subscribe / Unsubscribe Enewsletters | Login | Register

Pencil Banner

Sleeping better at an Indonesian hotel chain

Zafar Anjum | July 3, 2013
Jules Brookfield, Vice President of Technology at Archipelago International, one of Indonesia’s fastest growing and most technically advanced hotel chains, describes the challenges faced in running their multiple websites.

Overall AWS has always had more features, more experience, less bundling, cheaper, and a more pay as you go approach than most competitors, which I appreciate. They are innovating at such as rapid pace, all while reducing their prices again and again as the market leader, that's a philosophy that makes me feel good that I've made the right decision in the long term. I've also begun embracing their APIs to automate more and more of our tasks, which have led to huge efficiency gains and fewer mistakes.

How did the AWS solution work for you?

It's been great, and we've been increasing our usage of AWS dramatically over the years. I first decided I wanted to reduce the load on our servers by serving some and then all of our static content from the Amazon S3. Then I wanted to use a CDN to improve the load time performance around the world of our websites since our customers are located all over the word; Amazon CloudFront appeared just around the same time I started playing with AWS, that was my first easy steps into AWS. 

After that I decided I would sleep much better at night if I ran my servers on the Amazon Elastic Compute Cloud (EC2) and Amazon Relational Database Service (RDS), I would no longer have to worry that one of my IT guys would accidentally make a mistake while repairing a RAID array due to a failed disk and thus force a lengthy and sometimes error prone full restore on a Friday night while I was out on a business trip for example.

It also made it super easy to set up a staging/development server which then gets snapshotted and launched as our production server, further increasing reliability by being able to fully test changes in the environment they would be run in before launching them.  After that we just started piling on the AWS products, load balancers to increase reliability by running a cluster of servers in different Availability Zones, Amazon Route53 for DNS resiliency and performance (as well as a more automatic process due to the ability to use an API to make DNS changes - which is fantastic), etc.

Regarding our public websites, the results of this setup have led to some impressive metrics, such as:

* Our website page load times have on a global average more than tripled in speed;
* Our website uptime has gone from around 95 percent to more than 99.9 percent;

On top of that we also encrypt and backup our critical documents to Amazon S3, taking advantage of versioning, through an automated process. I'm testing Amazon Glacier now as well - although still waiting until it's available in Singapore Region before I really start using it. We also have internal applications which collect tons of statistics from our hotels, which run on a combination of Amazon EC2, RDS, DynamoDB, S3, and of course Route53.


Previous Page  1  2  3  4  5  6  Next Page 

Sign up for Computerworld eNewsletters.