Your company has developed a complex numerical calculation system used by some scientific and medical organizations in your country.
The whole application is using the Hadoop engine with an HBase database.
Because of the complex nature of the business, the overall setup runs in a private data center spread into two facilities for redundancy.
The application needs high I/O bandwidth for the workload.
Although they are not running the workloads with full capacity at the moment, there are some new developments in the business, and they are extending their services to a few other organizations in a different country.
Due to compliance reasons, they plan to host their application in the cloud instead of setting up data center facilities in the new country.
They want the same setup to be migrated over the cloud and still achieve the low latency, high I/O, and redundant cluster setup.
Please select a valid option that can help replicate the setup on AWS.
Click on the arrows to vote for the correct answer
A. B. C. D.Correct Answer: C.
Options A and B are INCORRECT because RAID 0 is used for the disk splitting, and RAID 1 is used for the disk level mirroring.
Because the instances might be placed on different hardware, the high-speed I/O may not be achieved as per the requirement.
Option C is CORRECT because the I3 instance supports the high-speed networking capability, and the Partition Placement Group can span across the multiple availability zones to achieve the required redundancy.
Option D is INCORRECT because the Cluster Placement Groups cannot span multiple availability zones or underlying hardware.
To replicate the existing Hadoop setup with HBase database in AWS, we need to consider factors like low latency, high I/O, and redundant cluster setup. Based on the provided options, the best approach would be to use I3 instances with Enhanced Networking enabled, as they offer high I/O performance and low latency.
The I3 instances are optimized for high I/O operations, and they come with locally attached NVMe-based SSD storage. Enhanced Networking is a feature that enables higher bandwidth, lower latency, and lower jitter network performance. It can provide up to 25 Gbps of network bandwidth and has improved packet-per-second (PPS) performance compared to the standard network interface.
Option A suggests using I3 instances with RAID 0 configuration to achieve disk mirroring. RAID 0 does not provide any redundancy, so if one disk fails, all data will be lost. Therefore, RAID 0 is not a suitable option for achieving redundant cluster setup.
Option B suggests using I3 instances with RAID 1 configuration. RAID 1 provides mirroring of data between two disks, which offers redundancy, but it does not distribute the data across multiple disks, resulting in lower I/O performance. Therefore, RAID 1 is not a suitable option for achieving high I/O performance.
Option C suggests using I3 instances with Enhanced Networking enabled and placing them in a Placement Group with Partitioned option to spread the clusters across multiple availability zones. Placement Groups are a feature that enables low latency, high-bandwidth networking between instances within a single availability zone. However, the Partitioned option does not spread the instances across multiple availability zones, making it a single point of failure. Therefore, Option C is not a suitable option for achieving redundant cluster setup.
Option D suggests using I3 instances with Enhanced Networking enabled and placing them in a Cluster Placement Group with an option to spread the clusters across multiple availability zones. Cluster Placement Groups are a feature that enables low latency, high-bandwidth networking between instances across multiple availability zones. It also ensures that instances are placed on different hardware to provide additional redundancy. Therefore, Option D is a suitable option for achieving low latency, high I/O, and redundant cluster setup.
In conclusion, Option D is the most suitable option to replicate the existing Hadoop setup with HBase database in AWS. It involves using I3 instances with Enhanced Networking enabled and placing them in a Cluster Placement Group with an option to spread the clusters across multiple availability zones.