- This topic has 3 replies, 1 voice, and was last updated 5 years, 7 months ago by
.
-
Posts
-
-
-
JBOD stands for just a bunch of disks.
It combines multiple drives (maybe of different sizes) as one volume for storage purpose.
pros: It is easily scalable i.e. we can add different types/sizes of hard drives .
unlike RAID, it avoids drive waste.10 GB +20 GB makes for 30 GB of total space unlike RAID wherein the net space will come out to be 20 GB.cons: It does not offer redundancy.
-
JBOD stands for Just a Bunch of Disk. As the name suggests it is just a storage technique where all the storage disc are mounted on a server/machine. The advantage here is these discs need not be identical and can be of different storage capacity. JBOD does not provide Fault tolerance(data redundancy) on its own neither it provides fast read/write operation as compared to RAID(Redundant Array of Independent Disks).
The only advantage is the storage capacity of each disk is fully utilized.
But then the question arises that why don’t we use RAID 0,1,5 etc instead of JBOD in Hadoop(Datanode to be specific).
The answer is very simple : They both complements each others very well. that is, all the short comings of JBOD like no Fault tolerance,slow read/write operation etc are taken care by Hadoop framework. And the major advantage that JBOD provides is, as stated earlier, it allows us to configure different disks of different make,sizes either in a single logical volume or a group of individual hard drives and utilizes these disks to their maximum capacity.In RAID disks are not utilized to their maximum capacity as usually one drive is kept for Parity information.
Another advantage of JBOD is if a disk fails in JBOD configuratioon , it would not affect HDFS while in RAID, if one disk fails whole array becomes unavailable. -
What is JBOD?
1. Just a bunch of disks. One of the storage architecture.
2. In this, All the disks are connected sequentially, and data write/read also happens in sequentially.Advantages
1. It is easy to scalable, when running out of memory,disks can be added on fly.
2. This will support different disk sizes. For example, you can add 200Gb as first disk, 300gb as second disk, 1Tb as third disk, 250Gb as fourth disk and so on. Same or Different volume can be used through out the architecture. But, at the end it will be configured as a single logical volume.
3. In case of any one disk failure, the data on the faulty disk will be lost. But, still data available on other disks for write/read.
4. This architecture is significantly cheaper compared to RAID.Dis-Advantages:
1. It does not offer redundancy.
2. This configuration lacks by performance compared to RAID(Redundant Array of Independent disks). Because, Read/write data happens in sequence, which takes more time.
3. No built in Data protection in JBOD configuration.JBOD in Hadoop?
1. Hadoop as a storage layer called HDFS(Hadoop Distributed file system) to storage huge volumes of data, where all the disks are configured with JBOD storage configuration.
2. Scalability:
As we know HDFS can scale on fly because the configuration will allow to disk expansion at any time without any limitation.
3. Fault Tolerance:
Hadoop is highly fault tolerant not because of the JBOD configuration, which doesn’t have data protection. One should take care of the data by the replication facility.
4. Cheaper:
JBOD configuration is implemented by using SATA drivers(commodity hardware), which are cheaper.
-
- You must be logged in to reply to this topic.
