Ben Stopford gave a really good talk on different approaches to scaling data systems at Jax Finance a couple of months ago: 'Elements of Scale: Composing and Scaling Data Platforms'. The transcript is available on his blog here: http://www.benstopford.com/2015/04/28/elements-of-scale-composing-and-scaling-data-platforms/

He gives an objective view of the various approaches for storing data in nearly every type of nosql store you can think of. Unlike most articles covering nosql he concentrates on the underlying storage mechanisms in each case rather than the high-level features of the platforms.

Stephen

On 21 July 2015 at 14:53, Jon Hatfield <[address removed]> wrote:

Hi all,

I am interested in scalability and HA because all systems have the potential to grow and also because the amount of data and internet usage always seems to be growing. I know about classic techniques like Apache load balancers (for both scaling up and HA) and using JMS for HA (by making use of clusters, queueing and asynchronicity) however this is no longer really cutting edge, does anyone know some good modern techniques? Or modern techniques to combine with these older techniques? I have used Zookeeper but not in depth enough to understand it well and have recently heard about Curator too which builds on Zookeeper, has anyone got experience using these technologies in a demanding clustered system?

There is also the data storage to consider in terms of how to distribute it and avoid lock timeouts. In my experience I have used master slave replication, writing only to the master node and reading from any node, which is pretty good when there are much less writes than reads going on. It would be great if anyone could share other successful ways to scale databases too.

One other thing - from a web perspective, when I last worked on the front end too over a year ago, there was the issue of sticky sessions to consider. Do you think this is becoming redundant as most systems are now stateless? For example, I have seen cookie IDs being saved in a database in a MongoDB course I did, avoiding the need for session state to maintain this kind of important info. Without session state there is no need to ensure that a user is always directed to the same node.

I hope some of you folks find this topic interesting too!

Cheers,

Jon

--
Please Note: If you hit "REPLY", your message will be sent to everyone on this mailing list ([address removed])
This message was sent by Jon Hatfield ([address removed]) from LJC - London Java Community.
To learn more about Jon Hatfield, visit his/her member profile
To report this message or block the sender, please click here
Set my mailing list to email me As they are sent | In one daily email | Don't send me mailing list messages

Meetup, POB 4668 #37895 NY NY USA 10163 | [address removed]

--
Please Note: If you hit "REPLY", your message will be sent to everyone on this mailing list ([address removed])
This message was sent by Stephen Henderson ([address removed]) from LJC - London Java Community.
To learn more about Stephen Henderson, visit his/her member profile
To report this message or block the sender, please click here
Set my mailing list to email me As they are sent | In one daily email | Don't send me mailing list messages

Meetup, POB 4668 #37895 NY NY USA 10163 | [address removed]