Ensuring reliability of SAN traffic over IP
Time outs are common problem for vendors designing their applications, when it comes to providing access to data connected to SAN’s. Network latency is another problem that faces the vendors .When such problems occur, it may affect the smooth functioning of the whole system. There is always a solution to every problem. So, many methods have been used to ensure reliability of SAN traffic over IP. Now a day, many IT organizations are using technologies like virtualization to reduce the cost of operation s, as well as to improve service reliability. The networks which support the virtualized storage area network s(SAN) is an important challenge that is facing many vendors, in fact, the protection of datacenter communication sis an important challenge to the network administrators.
In order to increase the reliability and performance over WAN and other networks, the end to end solution must be able to transport the storage traffic. An IP – San technology has been developed in order to maintain wire speed through put off storage traffic (Clark, 2003). This technology is used for connecting storage across optical and IP infrastructures. Within the datacenter, SCSI is relied for the data to be delivered promptly and safely. In fact A Transmission Control protocol controls data recovery and congestion. But the main problem with this is reduced performance. On the contrary, a user datagram protocol (UDP) can deliver superior performance. The best part of it is s connection less protocol. So; it is left to The IP- SAN connection to ensure the reliability of the storage traffic.
Let us first have a look at what San and IP are?
SAN is otherwise called storage area network. It is in fact SAN connects the remote computer storage devices are attached to servers. It is with the help of SAN, the devices like tape libraries, disk arrays and juke boxes appear to be connected to server. It is really the file systems that are built on SAN which helps in abstracting the files. With the help of SAN’ storage capacity utilization can be increased. Another plus point of San is that one can get access to data, that otherwise requires high speed block level access .Fiber channel has been one of the prominent SAN architecture over for many years. Due to high allocation costs, many people have been opting for some low cost SAN’s, now days. The earlier SAN had a humble beginning. In the past, they were used to handle small scales of enterprise systems. These systems, inturn were connected to arrays of storage. Present day San’s have become critical IT component.
Some important protocols of SAN include FICON, FCoE, FCIP, Iscsi. Let us see what these terms mean. Ficon is in fact fibre connectivity. FCoE is fibre over Ethernet. iScsi is IP over small system interface.
The benefits of SAN include shared storage as well as promises shared data .It uses faster SCSi speed and lower SCSI overhead. The storage management is easy and few operators are needed for more storage. Control is centralized and there is across platform virtualization.
The IP is known as Internet protocol. The network using an IP is called IP network. This internet protocol uses a suit called IP suit which is popularly known as TCP/IP (Blank, 2002) .This is mainly used for communication of data cross a packet switched internetwork. The main function of IP is to deliver packets of data from the source host to the destination host. The packets of data are otherwise known as datagrams.this delivery of data takes place according to the addresses popularly known as IP as address. This is just like our friendly neighborhood postman delivering a letter from one person to another depending on the receiver’s address. The Internet protocol uses certain addressing methods and structures for data encapsulation. The most widely used versions are Internet protocol version4 and version 6.The IP is still the predominant means of communication among enterprise and service provider market segments.
Let us take a close look at network latency .All of us know that bandwidth determines the speed of the network. Latency is the other factor which determines the speed of a network. In fact it is the latency which causes delays in delivering data from a source host to destination host. Depending on the delays caused latency is termed as low latency and high latency. Low latency results in small delays, while high latency results in longer delays. High latencies affect the actual bandwidth .Procession delays and transmission delays are some other delays which involve network latency. Ping tests and trace route are methods used to measure network latencies.
Some methods used to overcome network latency are shared database, maintained data copies and file transfer .In the case of shared database, the application which a person I s integrating read data from the same database. Another way is to maintain copies of application’s database .This will help the other applications to read the database. While using file transfer, one should transport a file which is an extract of application’s database. In this way other applications scan load data from other files.
Internet SCSI uses certain SCSI commands and storage traffic over existing IP networks (Jepsen, 2003). These are some ways in which SAN problems are tackled .One important way is to transport SCSI over IP.This has many advantages including easier maintenances increased system reliability, and of course lower costs.
To provide solutions to issues like business continuance and disaster recovery, An SL-IP 1000 SAN gateway has been devised by the SAN Valley systems. This provides seamless connection across WAN and MAN networks. In these networks, strict controls are placed so that issues like packet loss, latency and congestion is limited.
Now let us analyze the methods which are used to ensure reliability of SAN traffic over IP so that applications relying on SANs do not time out
Dark fiber or DWDM can be used to increase reliability, performance nd high security (Lovink, 2003). But high costs are always a threat in this. This problem can be tackled using a SL 1000 to the network, by using such an option superior performance, reliability, and security can be guaranteed. Internet protocol network are available in different speeds and qualities. But as the public internet cannot guarantee quality of service, it cannot be depended upon. Therefore, the best and most effective method is to use managed IP network connections. Such a connection can provide both the scalability of a packet switched infrastructure as well as reliability of a circuit switched network.
A network should not only maintain a superior performance, but also use a flow control n mechanism. With the help of a flow control mechanisms, the problems of congestion and latency can be reduced. Rate shaping is another method which can be used to ensure the reliability of San traffic over IP. The rate shaping is done by sending just enough buffer credits to SAN. For example, if a network can maintain only a lesser amount of MBPS, the rate shaping algorithm will send only that much mbps on the network. The SCSI handles all the retransmissions. This is done with out affecting the performance of the storage application and without windowing.UDP is effective mechanism which continue wire speed.
The SL1000 is an effective method which can be used to ensure the reliability of SAN traffic over IP.With this mechanism, an effective flow control, data recovery and traffic shaping is obtained .the biggest plus point of this mechanism is that it is able to deliver wire speed storage devices t long distances. Again, there is zero impact on the reliability of delivered data.
Another method is to use ISISI SANS which is beneficial to the people who use it in terms of labor, cost and reach (Desnoiresterres, 2009).The FC SANS are termed as small islands of technology inside datacenter. With the help of IP networks, the ISISI can get a global reach. This means that data can be accessed from anywhere, any part of the world.
Some other methods that can be used are push versus pull, synchronization logic, granulity etc for ensuring reliability .Granularity means getting a greater chunk of information at a single time. This will really help in getting the files safely to the destination source.
Since San is an important factor in IT industry present day SANs are designed to prevent overloading. An important solution is to build SANs which can accommodate ports for receiving and sending data at full speed.
Present day sans are designed in such away to prevent frame loss and network latency it is also designed to combat delay by integrating non blocking switch fabrics. The modern SANs work on the assumption that all hosts work simultaneously at the same time. The present day SANs are designed, keeping in mind a fan out ratio of 6:1and 12:1between host connections and array ports of storage .this ratio vary depending upon the storage activity. Hosts which need low utilization can use low fan out ratio, while the higher utilization hosts need a higher fan out ratio.
An ideal method to ensure reliability of SANs over IP is to design a good SAN. Now what does a good SAN design mean? A good San should be able to accommodate high intensity hosts as well as low intensity hosts at the same level. The San should be managed well and effectively.
ISCSI protocol has evolved another method called iSCSI boot .this allows the administrator to boot server operating systems over an iScsi based SAN. This in turn helps to simplify server management and creation. The distribution and maintenance of server images is also done with such a function. Prominent players in computers namely Broadcomm and Dell have come up with latest technologies to enhance iSCSI boot functionality on Broadcomm Ethernet controllers. The Use of SANs helped in block level storage over a network. This helps to increase the utilization of storage and at the same time reducing the number of devices for backup and manage. The Iscsi helps in simply booting an Os from a network. This allows the use of standard Os disk facilities. As the demand for SANs grow, importance. The network boot solutions like iSCSi boot also increases. Many enterprise data centers have started using them for effective management.
Some of the benefits of using SAN s are effective management; sudden disaster recovery etc. in the case of effective management, consolidation is an important fact. Consolidation helps to reduce the total ownership costs, creation of an effective data centre. Other benefits include retargeting of servers.
Enterasys has also come up with a solution to ensure the reliability of San traffic over IP. The Enterasys provides an identity based approach which ensures the reliability of SAN over IP. It optimizes the iSCSi communications through automatic discovery and prioritization of iSCSI SANS (Worden, 2004).
All these methods can be used to ensure the reliability of SAN traffic over IP.
Blank, Andrew G, TCP/IP jumpstart: Internet protocol basics, 2nd illustrated (2002), John Wiley and Sons, ISBN 0782141013, 9780782141016
Clark ,Tom, Designing storage area networks: a practical reference for implementing fibre channel and IP SANs, 2nd illustrated (2003) , Addison-Wesley, ISBN 0321136500, 9780321136503
Desnoiresterres, Gustave, Les Cours Galantes, illustrated (2009), BiblioBazaar, LLC, ISBN 1113110996, 9781113110992
Jepsen, Thomas C, Distributed storage networks: architecture, protocols and management, illustrated (2003), John Wiley and Sons, ISBN 0470850205, 9780470850206
Lovink, Geert, Dark Fiber: Tracking Critical Internet Culture, illustrated (2003), MIT Press, ISBN 0262621800, 9780262621809
Worden, Daniel, Storage networks, illustrated (2004), Apress, ISBN 1590592980, 9781590592984