Some of the forums users were also concerned about the cost of TSM. What is the cost structure?
And, how does this stack up with the competition?
TSM is priced on a per processor basis allowing customers to scale up as their storage needs grow. By allowing customers' storage resources to grow with their business TSM can actually save businesses thousands of dollars in software licenses. TSM also has many base features that its competitors charge for. For instance, a web administrator facility, encryption, off-site copies,
A disk pool is the "term" we use to describe a pool of space that resides on disk that can be used to store customers' backups and archives. The advantage of a disk pool is it allows multiple customers to back up and restore at the same time. Since it is disk it takes advantage of disk's random access nature. Since it is disk, there is no mount wait time or no tape positioning time when writing or reading data. With TSM, can you do both file level and volume level backups? Does its database keep track of which tapes belong to which servers?
Yes. TSM can do both file level and volume level backups. TSM tracks where every single file is. It also tracks where copies of the primary backups are. If there is a tape failure on the primary copy of a backup tape, TSM automatically utilizes the copy of that tape. If the copy of the tape is off-site it will request the tape be brought back on-site. What are progressive incrementals? Does this fit into TSM?
IBM defines progressive incremental as a backup cycle that only backs up new or changed data eliminating the need to ever do a full backup, and eliminating the overhead of full backups and their dependent incremental ordifferential backups. TSM offers customers progressive incremental backup, while competitors still offer their customers full plus incremental/differential backups. Progressive backup reduces the amount of data that has to be sent, thus requiring less network bandwidth, less time, and less storage. Most importantly, progressive backup results in faster restores because less data has to be restored. In one of our forum posts a user wrote this, "TSM and the majority of D2D solutions are just temporary band aids. They usually write data to disk in a tape format (serially). They create a staging area and another level of complexity." He also says users should be looking at a solution that takes advantage of the random access nature of disk and some take advantage of parallel read/write capabilities of many disks working in unison. How do you respond to these comments?
This is incorrect. TSM writes to disk pools that take advantage of the random access of disk. It allows for parallel read/write. It allows for multiple users to be accessing the disk at the same time. The disk pools can be used as staging areas, or can be used as the final resting place for the data. There has been a lot of talk lately about ATA arrays and backing up to disk. Does Tivoli Storage Manager facilitate this and if so, what are the benefits?
IBM Tivoli Storage Manager has been doing disk-to-disk backup since its creation 10 years ago. The benefits of doing disk-to-disk backup is that multiple machines can be backing up to disk at the same time. Unlike with tape, where either only one client backs up at a time, or multiplexing has to be implemented which fragments the data and causes restores to be very slow. Likewise, restores from disk pools are fast because there is no need to mount and position tapes, and once again multiple machines can be restoring from disk at the same time. IBM Tivoli Storage Manager is able to further capitalize on disk pool backups because it can migrate the data in the disk pool automatically to other types of medias. When it goes to migrate it is able to group all of a user's data together, so it is written in a way to tape that makes restores even faster.