Community Article #4

My name is Navid Malek Ghaini.

I was born in Tehran, Iran. Iran is a very beautiful country, especially from nature and historical perspective (have a look). I’ve recently graduated from the Sharif University of Technology which is the highest-ranking university of the nation and also recognized internationally (most of the graduates apply for the top 50 universities around the world). My major was computer engineering; however, it is called engineering but it is rather computer science (in Iran CS and CE are relatively the same). Currently, I’m a research fellow at INL-LAB and we are working on encrypted traffic analysis; I’m mainly at the network and OS side.

My mother-tongue language is Farsi (aka Persian), I’m also fluent in English. My preference for coding languages that I use is Python, C, JavaScript, and BASH (if considered as a programming language!).

About a year ago, I was a DevOps engineer in a company based in Iran called Nopayar; I was part of the project Yumcoder where we developed a scalable and extensible cross-platform infrastructure with the help of open source technologies. Obviously, at some point in the project, we needed a distributed network file system/storage. I considered so many available options back then, but in the end, I’ve chosen LizardFS for my purpose. The other famous options which I considered were Ceph, BeeGFS, GlusterFS, and ZFS.

I considered LizardFS to be the best open-source solution that existed. The main advantage of LizardFS was its simplicity in comparison with its other open-source rivals. The configuring and deploying of every node and even the whole cluster of nodes is not only easy but also relatively fast. The web GUI for cluster management is also a great benefit that made the work much faster and easier (especially for team leaders and managers who are interested in GUI and integrated information about nodes and cluster).
We needed an open-source platform to further configure and even alter it for our scalable infrastructure. My main interaction with LizardFS was to scale and tune the performance of it, and I should say that I didn’t have any major problem paving this path; the parallel file serving system was extremely well coded, flexible and easy to both tune and alter. Also, the community and the developers were both fast responding and friendly. Moreover, the replicating strategies were not only quite easy to configure but also highly advanced; for instance, we used a slightly modified version of EC replication of LizardFS which was one of the most advanced replicating strategies among open-source distributed file storages back then.

I had been using LizardFS for around four months and it was a great experience. However, it had some minor performance issues with SSD storage back then (which are now history), but I believe it was by far the best option available.

The statement on LizardFS’s website “Get Your Storage Up and Running in 28 Minutes” is the greatest thing about it. Plus, LizardFS is highly configurable and customizable and indeed fast and easy to scale.
With the new team and development strategy, I believe the minor issues that it had, like the one I mentioned, will be eradicated and it will become one of the best-known options, if not the best, the network defined storage in the near future.

In the end, I would like to thank Mark Mulrainey for starting this exciting movement of sharing the community experiences with LizardFS.