Discussion on cloud integration and large data

Integration is central to today's it's main disruptive forces. Searchsoa.com, editor of the Web site, Jack Vaughan, recently http://www.aliyun.com/zixun/aggregation/16482.html with the Blue Mountain Laboratory >david Linthicum a conversation, and they talked about three key topics: integration, cloud, and big data. The scope of this discussion ranges from cloud integration to cloud security issues and the application of Hadoop.

In particular, linthicum the best practices of the cluster, including how companies decide which services are best for the cloud. The core of his guidance is the focus on assessing existing enterprise systems.

My advice is to look at your architectural concepts-your existing business. According to their systems, divide them into their component parts and then decompose them into the core services provided by the system. Then, depending on where they are placed, they personally evaluate those services.

To also decide that services are placed on the local cloud, Linthicum suggests looking at the role of service, how much computing power they need, and what kind of information they can digest and produce. He goes on to say that in some cases the organization can mix and match the cloud and the local system. He says this approach will provide more functionality to agile businesses.

According to Linthicum, the needs and challenges of the cluster depend on the size of the enterprise. However, large enterprises have many systems and related large budgets, but small or newly created businesses usually have limited resources and fewer systems.

Large enterprises can require more complex integration projects, but they also face difficult, time-sensitive issues when it comes to integrating strategies and enabling technologies. But for small businesses this is different, Linthicum says:

"Small businesses" usually do very simple integration scenarios. I don't think they're going to involve very complex, event-driven, service-oriented stuff ... This is because of the relationship between expenditure and complexity.

Data integration in the cloud

Most cloud discussions are followed by security issues, which is understandable. More data are being recorded than ever before, and companies, big or small, are hesitant to move their confidential information from the local area. On this topic, Linthicum gives advice, some of which may be surprising:

I think people tend to overestimate the amount of security they really need. ...... Realistically, because security will cost you money. A lower level of security is available when you protect unwanted data that is not necessary and does not cause you harm, or if they are stolen in some way.

When it comes to storing and managing a lot of information, Linthicum says, big Data systems are dwarfed by their peers, object databases, and XML databases. Large data systems like Hadoop are a way of managing complex, explosive unstructured data today, he explains.

They don't need any structure to happen. You can have a huge amount of information in a file--something that we wouldn't call a database for many years ago, and you have the ability to map it to a struct in analysis, so it's easy to understand.

With cloud computing, companies can now take advantage of Hadoop because of its ease of access to commodity processing, he continued. Today, according to Linthicum, this is the only answer to the huge business problem of integrating large data and cloud data.

(Responsible editor: The good of the Legacy)