Smart devices

We all use some electronic devices, equipments and gadgets. Have you ever tought how do those work? Most reliable devices are simple. Your doorbell does not contain “intelligence”, it does not consume much energy and does that job very well. Unless it is broken physically. You don’t need to update software on your doorbell, you just may need to replace battery. Which probably has run more than 5 years. Same is with your alarm clock next to your bed. Old technology, proven reliable. Maybe it may go a bit advance in a month but mostly it does job very well. No need to connect internet nor update security patches.

Best and most reliable devices are not trying to be smart or intelligent. They just do the job.

But – we are going to world where devices try to be smart. And we are allowed to be dumber. Is it right direction? Can we produce more with intelligent devices around us? Are we more creative? Do we need all that information offered to us?

Modern storage solutions

We all have computers. It is either traditional desktop computer, laptop, tablet or a mobile phone. And all of those require some kind of storage. At least it is used to read and start operating system from there. But it is also used to keep photos, documents and temporary internet files. Mostly data storage is inside your device. But there’s even more storage available and used at internet. That’s where you stream video content from. Or upload latest travel pictures to be shared with friends.

But what storage options there are available? On internet you can have free data storage from multiple vendors like mentioning few: Google Drive, Dropbox, OneDrive, Degoo and Mega.

But where all those systems actually store their data? On a hard disk drives you may know. And your file is not only on one hard disk drive but spread into multiple drives on multiple computers. This system makes your files to be always safe and quickly available. If  your storage service provider has a broken disk drive you won’t loose your file. You won’t miss your files even if that server computer breaks where your files were stored.

Technology there is available for everyone willing to have such services. If you are IT service provider you may have such storage system already. It is either based on well known storage solution provider. Or it may be built on regular servers. Unfortunately traditional storage system providers have hardware which requires special disks for it. And it typically has own limitations – either hardware is getting old, slow and there is no changes to expand it beyond limits. Special disks are required and prices might be very high.

But there is a better option – well known, proven and everywhere used but never before combined like it is possible nowadays. Use regular server disks. Those are cheaper, not restricted to be from certain manufacturer,  can be swapped to larger capacity and configured quite freely. And when clustering multiple servers together those can provide redundant, fast and highly available disk storage systems. But – there you also would need to have a server software which manages all data replication, distribution and managing fault tolerance.

There Red Hat has two options on their Red Hat Storage platform. Either Gluster or Ceph – depending your needs. Both of these are based on open-source-software which are also available freely but without any support. Generally these would scale every direction (up and out) and almost infinitely. If you need more capacity you can buy more disks to existing servers or buy more servers. Or if your environment is not quick enough you can replace slowest systems with newer and faster servers. And do all this without any interruption. These systems would share storage via multiple protocols over IP. Most traditional is CIFS (Windows share) and NFS but also available via iSCSI, Object Store and native protocols (like GlusterFS and CephFS).

Red Hat Gluster would run on minimum of two hosts. It can replicate data but also distribute it between hosts. Data replication would keep it redundant but data distribution would make it accessible faster. So then you could run data from multiple hosts at same time. Pretty much like RAID levels 0 and 1 on local host but this would be over network.

Red Hat Ceph is for more complex environments. It spreads all data as a junks to storage servers. Data is spread and replicated by using CRUSH map. And with Mon servers data is always replicated elsewhere if any data brick is lost because of hardware issues. This storage mechanism is suitable for huge amount of data like for retrieving data from nuclear plant sensors and used on scientific calculations.