SMI-S expect more

SMI-S was originally intended to manage heterogeneous environments on a single console, and we have reason to look forward to it as it develops.

Why SMI-S?

What users do not want to do most is to be bound by the vendor. No matter how well the current technology of this product is, the products of any manufacturer cannot guarantee the long-term mainstream in the market. For example, my favorite laptop brand a year ago isIBMAlthough the brand is still present, the owner of ThinkPad has changed.

In the Ethernet field, the interconnection problem between switches is solved very well. However, the situation in the fiber channel field is quite different. If all Fiber Channel products comply with unified management specifications and can communicate smoothly, users do not have to worry about vendor binding.

Storage Management Initiative specifiCATion (SMI-S) is expected to change the chaos in the San management field. Before that, San users must install various management software designed by different vendors according to their management interface specifications to use devices of different vendors.

The advantages of developing a general management standard are obvious, especially in heterogeneous computing environments where storage products come from multiple vendors. Fundamentally, using SMI-S eliminates the need to learn new management tools or processes when adding components from different vendors. The SMI-S manages network components in the same way as other components on the network from a single console.

For manufacturers, the widespread adoption of SMI-S in the industry will have a profound impact. Currently, developers in the storage management application field are faced with the problem that they must concentrate on integrating interfaces that are incompatible with multiple device vendors. Various manufacturers are very different in terms of protocols, transmission technology, security mechanisms, and other fields, making integration more difficult. They all lead to slow release of new products.

Relatively speaking, the SMI-S enables storage vendors to use a single standard interface for independent development. Because it is no longer affected by the compatibility of multiple manufacturers' products, manufacturers do not need to develop and integrate heterogeneous and proprietary interfaces, thus shortening product time-to-market and more effectively launching and updating products, it also improves product performance.

SMI-S is based on some existing standards, such as Common Information Model (CIM) and Web-Based Enterprise Management (WBEM. The former describes the management requirements and capabilities of the system, and the latter specifies the use method. SMI-S includes interoperable and scalable universal management transmission, automatic discovery, and resource locking. All of this means that when a new SMI-S component joins a SAN, it will announce who it is and what it can do, and then it can share its resources with other components.

Industry challenges

The biggest contribution of snia is to bring together highly competitive vendors and work toward the bright future of "interconnection. Currently, the most specific score is SMI-s. However, the author found that SMI-S still has a huge deficiency, that is, it only in the "discovery" to do some passive work, active management work is still difficult to achieve. In addition, during the "discovery" application process, there may be situations where the vendor does not cooperate. Some point of view that the development of SMI-S in the current stage has encountered a lot of difficulties.

At present, the industry has more and more questions on SMI-S standards. First, some people think that it does not actually involve the hardware compatibility of devices provided by different vendors. Second, although the future version of the specification promises to be resolved, it does not involve problems such as backup, remote replication, device performance management, and storage management software interoperability. Another question is whether suppliers are willing to put standards into practice because they are more willing to provide different technologies to protect the market. Of course, this SMI-S will allow suppliers to provide their own extended interfaces for the purpose of extended functionality.

According to the original intention of the SMI-S, it can manage all the user's storage resources, but the current situation, it is still in the early stages of development: it can only find a variety of problems, but cannot take the initiative to manage. Discovery functions are of course important for implementation management, but it is only the first step of management, and the problem of SMI-S is that discovery functions are basically all of it.

In addition, the currently authenticated SMI-S Management Console only discovers HbA (host bus adapter), fiber channel network, storage switch, and arrays. Tape drives, tape libraries, optical storage, and other key components connected to San have not yet been discovered.

So how long does it take to implement it? It is hard to say at present. We can discuss this situation through the tape library example.

One vendor provides a tape library Management Console that can satisfy the SMI-S, of course, it can only provide discovery functionality, but has not yet been certified by snia. ADIC's iplatform tape library has built-in SMI-S capabilities, so other management consoles that meet the SMI-S can also discover them. So why is there no conformity test? The reason is that snia currently requires at least two or more vendors to submit a test request so that snia will schedule the test. This is easy to understand. Since the main purpose is to test the interoperability consistency test, it is clear that the tested object should be larger than a vendor. In terms of tape library management, no other vendors have launched similar solutions except ADIC.

Many tape libraries provide more and more functions, so the demand for management is also surging. Many of these functions are provided for specific products, and manufacturers hope to distinguish them from competitors. On the one hand, the vendor does not want these management functions to be added to the SMI-S, on the other hand, because other vendors cannot provide similar functions, it simply cannot join the snia interoperability test.

So what should a manager do? In order to gain a thorough understanding and management of storage, he still cannot give up on standard-based management. For example, ADIC, in addition to claiming that its tape library can be managed in the SMI-S-compliant console, it also provides APIs (application interfaces ), other vendors can use these APIs to manage the ADIC tape library. Currently, EMC uses this method: In the same EMC San topology, the ADIC tape library is displayed in real time together with other San components, however, the specific management work is still implemented through the API.

In addition, SMI-S also missed a very important management function, that is, remote storage management. Obviously, you need to manage remote data links, whether for Remote Disaster Tolerance or a remote device with a super-large San. However, if you cannot monitor the entire link of the data, you cannot understand every event of remote communication. Unfortunately, SMI-S has nothing to do with this. If users have management requirements in this regard, they can seek management software like Magellan of centrepath.