Will cloud computing age become the Terminator of fault-tolerant servers?

It has been farewell to the era of blossoming, steam. Once upon a time, Wang, DEC, Compaq, SUN, SGI and other such big business has faded out of sight. But strangely enough, Stratus, the representative of this fault-tolerant machine, is still alive and well. Since the birth of the 1980 Stratus, decades have passed, no more, but fault-tolerant server is always active.

Stratus didn't have a very, very, very stylish company style. Stratus is a typical technology company and is not good at marketing. In foreign countries, Stratus has a loyal customer base, the user's technical level is relatively high, do not need too much market cultivation and guidance. As long as users need high reliability, they will adopt Stratus or other branded fault-tolerant servers. Foreign users will not be entangled in technical issues. In contrast, the domestic market, although rapid development, but the market needs to nurture and guide. This is precisely not Stratus's strength.

Software vendors often do not actively recommend fault-tolerant technologies in terms of business interests. If the use of fault-tolerant server, whether the operating system, database, or application software license sales volume, compared to dual-machine redundancy to reduce one-fold. Therefore, some people joked that the dual-machine redundancy scheme is the choice of manufacturers, fault-tolerant scheme is the user's choice. If users insist on fault tolerance, software vendors will support it, but software vendors will not actively promote fault-tolerant servers. So is Microsoft, Oracle.

Technically, a fault-tolerant server can correct hardware failure to the maximum extent with the help of "lock-step technology". In the fault-tolerant server design, CPU, memory, I/O, and so on, all using redundant design, and through the lock-step technology, each calculation process for comparison, accurate to enter the next step. So the reliability of the system can reach 5 9 levels.

But with the advent of the cloud computing era, system reliability is all dependent on the cloud. Cloud computing through computing resource pooling, so that computing resources do not rely on the high reliability of a single system, through the cluster technology to ensure the overall high reliability of the system. So is the cloud computing era a terminator for fault-tolerant servers?

The answer is no. For many. COM enterprises, especially the use of decentralized deployment structure, not because of a single server failure, affect the system access. In fact, in the process of accessing the site, every query and access, even the same site, provide a background server is probably not the same, so, only access is assigned to the failed server, users will feel. But not all applications have such characteristics.

High reliability servers are still required for applications that require high reliability. Virtualization does not solve high reliability challenges. Take VMware for example, high reliability requirements need to purchase fault tolerance module, for low requirements of the application, need to purchase HA module. Among them, Fault tolerance similar to fault-tolerant, ha analogy to dual-machine redundancy scheme.

Does Fault tolerance have high reliability for fault-tolerant servers? It's too early to talk. For cloud computing, cloud computing also needs fault-tolerant servers as an IaaS platform.

(Responsible editor: The good of the Legacy)