Distributed file systems.

The drivers of choice. Part II. SAN to NAS In the first part reviewed disk storage systems, their organization, differences and choices. Today, many businesses need to create an organized storage and data management. Minimizing investment and the desire to improve the investment led to such many issues related to the objectives which they chose, and especially those technologies. Moreover, companies receiving and adopting these technologies, and without detailed answers to their questions in the planning process, have to invest additional investment to bring solutions to an acceptable level. In this part, we will try to define these two technologies and to answer the questions : What are they really? What they differ? What are the challenges in implementing these technologies? NAS NAS (Network Attached Storage) device has many characteristics of classical file server. They include highly efficient interface typically 100MB / s Ethernet and Gigabit Ethernet. Usually FC NAS use for internal communications between its integrated storage, and external communications with customers using Ethernet. Indeed, NAS is a hardware architecture designed for delivery of its unified storage, as efficiently as possible. NAS device is extremely easy to set up and configure. Most of them will be fully ready for use within minutes of installation. Users configure and manage these devices remotely, such as through a Web browser or a device supplied by the manufacturer. NAS devices. System-NAS devices simple and clear end-user. While they work on the universal operating systems such as Linux, Windows or less popular, require no special knowledge to customize OS NAS- device. In other words NAS- device is nothing like a normal file server that supports multiple network file systems and additional services with no extra features and added funds for quick setup via a web interface or with a simple menu on the front. When connecting to the network, NAS device replaces standard file servers. Typically, network resources are exported as NFS, SMB or CIFS file system. Also commonly used access through HTTP and FTP protocols To file systems, delivered to the network NAS- device can be used to directly access, via computer or by the client application server. Many of the server and client can simultaneously use the system storage. There is a choice-NAS devices. Prices are not great system and the average ranges from $ 1,000 to $ 100,000, they are targeted for small and medium-sized businesses. Large NAS systems (price over $ 100,000) are oriented to large industrial applications. The advantages and disadvantages of NAS The main advantages of NAS file service excellence, the ability to use different file protocols for various user operating environment, snapshot feature to quickly file a lower cost per gigabyte, and lower administrative costs. NAS significantly reduces TCO, using consolidated file servers or DAS (Direct Attach System) systems. One of the factors limiting the use of NAS is scalability. While capacity storage systems using the drive with a larger capacity, increased to tens of terabytes, the fact that the autonomous NAS products are for administering NAS devices directly attached storage system, limiting their capacity. To solve this problem, some manufacturers provide management device in a separate NAS functionality. While NAS has many beneficial sides Unlike conventional file servers, for the most part, it is not set up to block access, as required by specific applications, such as Microsoft Exchange 2000. While several NAS products can service applications to databases is still a problem of productivity due to the architecture of TCP / IP. A further conversion, information from the disk converted to the format of a network file system, a customer of a network system in a format already understood client OS. There is a double is not always necessary changes, which limits capacity as a whole, employs processor time and uses more RAM. Another issue of concern to some users is the lack of flexibility NAS systems, as they are designed as a device to certain fixed. While manufacturers now offer simultaneous file access and block access, and can join Manager NAS device to different types of storage arrays (SCSI or FC) or SAN structure. SAN SAN (Storage Area Networks) is based on a model connected to the server through a separate network storage devices, independent of the network organization and communication networks. Network storage systems can include disk drives, RAID devices, tape libraries and other equipment. Multiple servers, and even customers may obtain access to the network storage. Today, the largest network of technology used for FC SAN is, but there is an initiative Internet SCSI (iSCSI), which is the use of TCP / IP / Ethernet networks instead of FC SAN. SAN based on FC technology efficiently move data without giving further strain on the network. The original and simple manner using branches FC arbitration logic (FC / AL) can be used to connect a limited number of devices in the ring. More complex SAN using concentrators (hubs), switches (switches) and routers (routers) for FC structures that may include a large number of remote devices A key principle of SAN includes unloading communications networks. Data transfer protocols, such as Ethernet and IP are substantial overhead costs, while FC technology to send large amounts of data with great efficiency. SAN ideology in the next-to move away from tying a specific OS file system to storage devices, a data transfer protocol, which provided support (encapsulation) of many popular data transmission interfaces (TCP / IP, ATM, SCSI). All devices have the same address, and several ways to connect to the SAN environment, and into conflict with the joint work of several servers from one such tape device. SAN simple in the end user's network, so customers' computers, and Web users are unaware of the existence of server SAN architecture. Pros and SAN The main deterrent to the use of server storage systems, is to limit the number of storage devices that can be added to any data server and with the increasing complexity of storage. SCSI architecture is not suitable for very large, capacious storage systems. Quantities devices per SCSI chain and the number of slots for the main bus adapter per server, is often not enough to reach many of the necessary total capacity system. FC-AL system can serve up to 126 devices, a system based on FC, is capable of receiving one million devices. SAN servers can be physically independent from their storage devices. Using connections with a SCSI cable, it requires that the devices are of each other at a distance of several meters, and using FC, the distance may be many miles away. This allows centered equipment for storage systems, even if the servers are scattered. Using SAN, it can create a large pool centered storage systems connected to the server, which can be placed almost anywhere. SAN enables organizations with large information centers, the distribution of access storage systems regardless of the physical situation. SAN may include tape libraries, which can be used for backup devices in the network. Among based only on the LAN backup operation can fill a network. A SAN provides the opportunity to book, without moving any data on the network and perhaps even having to log into a server. On the other hand SAN has its minuses, it is expensive and confusing. IT departments need enough qualified staff to manage the SAN. Fiber-optic networks and communications equipment necessary for a SAN also require significant investment. Still, for many of the overhead costs to implement and maintain SAN less onerous than the problems associated with maintaining the ever-expanding file servers. One of the main points that is worth drawing attention with the introduction of a SAN reliability and fault tolerance. Strong SAN includes many redundant components. In the internal disk is configured RAID which continues to work even if one or more drive out of service. Many controllers and power supplies also increased reliability. Another time, which limits the use SAN that it is impossible for many servers with different OS join one device for direct access to files. But, the fact that the SAN is SCSI protocol and block access to the data needed investment in additional equipment and critical task of sharing in the SAN network (cache synchronization of file systems on servers). Examples of such programs can be Tivoli SANergy. Looking at the steadily declining prices disc media may be giving the wrong impression about the cost of managing storage systems based on SAN or NAS. Especially in the SAN environment, but to some extent and NAS, the price of storage is much higher than on disk drives. The switches, high performance controllers, control software, the internal processors and other devices as they improve performance, reliability, scalability and manageability, and increase the overall cost. ANALYSIS Both network storage technologies SAN and NAS provide a central data storage, meeting with various tasks. NAS is already proven technology that supports most hardware storage and network equipment, and providing Internet access to files for users and applications through protocols NFS and CIFS. Data transmitted over a common network, which increases the pressure and impact on productivity. Consequently NAS technology is a good choice for : - Local heterogeneous networks, which employ mostly UNIX and Windows servers. - Access to data through Ethernet or global networks. - Above data is not performed operations requiring high capacity subsystems, I / O (such as reading files). Main applications are Web access, shared file resources, centralized storage of customer data. On the other hand SAN already passed the initial phase of its evolution, and now offers more potential amenities than NAS. Sitting in servers, SAN is a private, more scalable and easily accessible system. Because of its high-speed connectivity SAN seen as a replacement for LAN backup and recovery of the external media. But still there were problems with the compatibility of a particular component SAN, access to the data in heterogeneous networks and services for network SAN. Application SAN technology can be more effective when : A high-security and availability of data and apply cluster solutions. - A consolidation of data in a secure environment for systems based on UNIX / NT / Novell - A centralized data to BD-scale enterprises - There is a need for a secure and high-capacity and data access at the level of blocks (for example, database or video) - A backup of large amounts of data in a short period of time without the burden of a standard network