14BigTable BigTable is a distributed storage systemthat can be used to manage 11structured data. It is usually designed to scale to thousands of machines and megabytes of data reliably. Big table can be used to achieve many goals that decreases the total cost of ownership and increases the returns on investment. These goals include scalability, wide applicability, high availability and high performance (Luo, 2010). BigTable are very useful in7demanding workloads which may range from throughput-oriented batch processing jobs to servicing latency-sensitive data to the end users. The clusters of the Big Table used are able to span o a wide range of configurations from few to thousands of servers.They are also able to11store up to hundreds of terabytes of data.This decreases the total cost of ownership considerably. Another advantage of the BigTable is that at6resembles a database and it shares many of implementation strategies with the database. Main memory3databases and parallel databases have achieved high performance and highscalability. The BigTable provides an interface that is different from6such systems. It does not support full relational data model but it provides theclients a simple model that can support dynamic control over the data format andlayout. It also enables the3clients to think about the properties of the represented data in the underlying storage. Clients are able to control the locality of data through making careful choices in the schemas. The BigTable schema parameters also allow the clientsto dynamically control on whether to serve data from disk or out of memory.This feature attracts, any users and the return on investment is achieved. Other advantages that reduce the total ownership costs include the fact that the joined operations will be less costly due to the denormalization. Replication of data performed by the BigTable5is less costly due to data independence. This means that if an individual wants to distribute or replicate data across two nodes, they will probably do not have the problem with having the entity one node and related entity in the other nodesince similar data is normally grouped.Dynamo In dynamo data is usually partitioned and distributed by use of consistent hashing and the object versioning facilitates12consistency. Dynamo uses the synthesis of available techniques to achieveavailability and scalability.1A quorum-like technique and a replica synchronization protocolare what maintain the consistency in the replicas during the updates. Also, dynamo employs a membership protocol and a gossip based replicated failure detection. This system reduces the total cost of ownership due to the minimal need of manual administration. Addition and removal of the storage nodes from the dynamo can take place without the requirement of manual redistribution or partitioning. This system has been used to buy companies such as Amazon, in its e-commerce as the underlying technology for important services. It has proved to be very efficient in scaling1to extreme peak loads and without any downtime even in the busy seasons. Dynamo is therefore capable of increasing the returns on investment due to increased production and strict performance demands. Dynamo is different1from the decentralized storage systems due to its target requirements. Dynamo is targeted mainly at the applications that require a data store that is always writeable and where updates are not rejected because of concurrent writes or failures. This makes it very productive. Dynamo is also built for infrastructures that are within a single domain of administration in which all domains are trusted. Another advantage that may reduce the total cost of ownership is the fact that1applications that use the dynamo do not require additional support from thehierarchical namespaces as it is with the normal file systems. It also does not require support from complex relational schema. Dynamo main advantages are how it provides is availability at significant levels. The applications are able to receive up to 99.99% successful responses and loss of data event is unlikely to occur (DeCandia, 2007). It also provides the necessary knobs using the N, R, W parameters to tune to tune into instant needs. Unlike other data stores, dynamo is capable of exposing1data consistency and the reconciliation of logic issues to its developers. This helps in reducing the total cost of ownership. The dynamo is also able handle various1failure modes and the inconsistencies that are likely to arise. Porting of these1applications to use the Dynamo is a very simple task. However for new applications, the development of the system will require some analysis so that they are able to merge with the business case suitably. Finally, this system is able to adopt a1full membership model whereby each node understands the hosted data by its peers. Each node in the system is able to share information with other nodes in the full routing table. Dynamo is therefore very advantageous in the increasing of the returns on investment and reduction of the total cost of production due to the provision of desired availability levels and performance. It is very successful in the handling of1data center failures, server failures and network partitions. It is alsoincrementally scalable and this allows the service owners to scale up or downdepending on their request load at the moment. The service owners can also be able1to customize their storage system based on their desired performance,consistency anddurability. The use of the Dynamo proves that we can combine decentralized techniques to generate a single system that is highly available. Definition of iSCSI Internet4small computer system interface is an internet protocol that is based on thestorage networking standards for the purpose of linking the data storage facilities that are developed by Internet Engineering Task Force. The iSCSI is used in facilitating data transfers over the intranets and managing data storage over long distances by carrying theSCSI commands over the long distances. This protocol is one of the key technologies that are expected to assist in the bringing about of rapid developments of the SAN market (Satran, 2004). This will be through the increase of the performance and capabilities of storage and data transmission. The IP networks are have the feature of ubiquity and thus the10iSCSI can be used in transmitting data over the Local area networks, wide area networks or over the internet. They can also enable retrieval and storage of the location -independentdata. Concept of TCP/IP network The IP2(Internet Protocol) addresses are the main forms of addressing the TCP/IP network universally. These network layer addresses normally identify each network interface uniquely and therefore serve as a2mechanism in which data can be routed to the appropriate network in the internet work. Both2the TCP/IP transport protocols use the concept of sockets and ports for thevirtual software addressing (Hunt, 2002). This enables2the function of numerous applications simultaneously on the IP device.By use of the sockets and the ports, the hard drives and tapes interface when attached on a single bus RAIDs. References8DeCandia, G., Hastorun, D., Jampani, M., Kakulapati, G., Lakshman, A., Pilchin, A., & Vogels, W. (2007, October). Dynamo: Amazon’s highly available key-value store. In ACM SIGOPS Operating Systems13Review (Vol. 41, No. 6, pp. 205-220). ACM.Luo, J. Z., JIN, J. H., SONG, A. B., & Dong, F. (2011). Cloud computing: architecture and key technologies. Journal of China Institute of Communications, 32(7), 3-21. Satran, J., & Meth, K. (2004). Internet small computer systems interface (iSCSI). Hunt, C. (2002). TCP/IP network administration (Vol. 2). “O’Reilly Media, Inc.".