Linux Enterprise Application Survey Report

Source: Internet
Author: User
Linux has been growing. This is an indisputable fact, but any growth is accompanied by troubles.

Linux has been developed by some open source code enthusiasts, and some people who are enthusiastic about the underlying technology research and use the open source Linux to build some unique applications on it, people hope that such an operating system can break some proprietary or monopolized systems. They feel that Linux should enter enterprise-level applications, especially key enterprise-level applications, in this way, we can get the real value of use and grow into a labor force that can be transformed into wealth. This is the reason why more and more system vendors that release commercial Linux versions have emerged. It is also the reason why more and more vendors in related fields have joined the campaign to accelerate their conversion to labor. Since the middle and late 1990s S, Linux system vendors and it giants are becoming increasingly enthusiastic about Linux, but what are the opinions of the users who make the payment to the labor force? What troubles have Linux experienced when it became a real labor force? I hope you can find some answers from this topic.

We recently conducted a small-scale user survey on our cnw website (www.cnw.com.cn). We believe that the survey results are of great reference value to users and provide answers to the above questions.

Part 1

Experience the cutting-edge of Linux applications

Question 1: How familiar is the Linux operating system?

Result data:

Figure: 36p17a.jpg

Data inspiration: users' familiarity with Linux is a very important issue. The degree of familiarity determines their acceptance in a certain sense. The results show that most users only "have access" to Linux, accounting for 37.5% of the valid survey results. The reporter learned from more in-depth phone calls that, this kind of "experience" includes some previous testing experience, including the experience of relevant customers, including some publicity and training sessions, etc, it is basically a very preliminary understanding, but it is not clear about some features and functions of Linux. Users familiar with Linux are generally R & D or maintenance personnel. Their daily work is to develop or maintain Linux-based Web servers and other applications around Linux.

Research Question 2: What do I think of Linux as an enterprise-level operating system?

Result data:

Figure: 36p17b.jpg

Data inspiration: in fact, in some small and medium-sized enterprises, Linux is a common phenomenon as an operating system for deploying web server applications, in our preliminary research on Web server testing based on application solutions organized by the testing lab of "online world" last year, we learned that nearly half of SMEs chose the Linux operating system for deploying web applications, this survey aims to understand users' opinions on using Linux as an enterprise-level key business application platform, so we have prepared four more specific alternative answers.

According to the survey results, no one believes that Linux is not mature in all aspects. Most users still think that the Linux system has complete and mature functions, but lacks application software. That is to say, lack of applications is the biggest drawback of Linux. Developers think that Linux is mature and available, and some integration engineers have helped customers deploy Linux-based applications. In an interview with hardware and middleware vendors that support the Linux platform, the reporter learned that the cost advantage mainly comes from the server hardware products in the IA architecture and the cost of the Linux operating system, users start to understand and accept Linux, but most of these applications are limited to the most mature web server applications of Linux applications. A few users started to use Linux-based databases and deploy application server software. However, compared with other system platforms, especially the Unix and Windows platforms that deploy key business applications, linux is still a minority. One of the reasons is that Linux is not mature, the system itself is not fully functional, and more importantly, there is a lack of applications. Unisys recently announced that its es7000 server can run Linux. Unisys cooperates with Novell and Red Hat to provide services and support based on Linux Solutions, they also announced the first dynamic Partition Function in Linux. Chen zhongyu, general manager of Unisys Greater China System Technology Department, said in an interview that although such cooperation is growing, there are more and more hardware platforms supporting Linux. However, compared with windows, there are still too few Linux applications, which takes time to work with more system vendors and ISVs. He believes that, it takes at least two to more time for Linux to enter key enterprise application fields in scale ~ 3 years.

In addition, in the answer to the question "Why I don't choose to deploy Linux", 62.5% of the respondents chose the lack of application software, applications are the key to promoting enterprise-level Linux applications.

Research Question 3: What kind of applications does the Linux system deploy?

Result data:

Figure: 36p18a.jpg

Data inspiration: In the answer to this question, most users choose to deploy Web servers. This proportion reaches 75%. It is believed that Linux can be used as a platform to deploy database servers and large-scale application software/key business applications, accounting for 22.5% of the platform. The proportion of users who think they can serve as database servers and deploy application server software platforms is also large. According to some surveys on the Internet, the Linux-based open-source database is getting more and more attention. The challenges to the market come from three main competitors: MySQL, MaxDB, and s SQL. In fact, deploying large database servers is also one of the key enterprise-level applications. From the data, users hope that the scope of tasks that Linux can undertake has been expanded and gradually approaching Key Fields, there are many reasons for the strong promotion of system manufacturers. Like Oracle, on the basis that all database products support the Linux system, it is still vigorously promoting the application and development of Linux, this includes kernel-level R & D and development of new database products for some commercial Linux systems. Although it has the motivation to break into the operating system with the opportunity of open source Linux, however, as one of the most important database vendors, Linux is still very effective. Unisys believes that the server products in the IA architecture can be combined with the flexibility and openness of the Linux operating system to provide users with new products with both performance and price advantages, in the field of high-end applications, we can replace the expensive, but also the high-end, server-based proprietary systems. Manager Chen zhongyu of Unisys told reporters that when it comes to the deer-hunting Unix/RISC market, the application field of the traditional RISC system should be the breakthrough point of Linux's direct import to high-end key business applications.

It is worth noting that users from insurance, power, and other major industries have not chosen to deploy Linux as a platform for running key services. In more in-depth phone interviews, they are conservative about this issue. On the one hand, they are not familiar with Linux, and on the other hand, they are still skeptical about the security and function completeness of Linux.

Research Question 4: What aspects do Linux, as the core enterprise application system, need to work hard?

Result data:

Figure: 36p18b.jpg

Data inspiration: From the data point of view, application software support is the most prominent problem. The following figure shows that the number of users who lack professional application software reaches 72.5%, and the number of users who lack large management software and application server software exceeds 50%, that is to say, professional software built on Linux for special applications is urgently needed by users. Independent Linux vendors should increase cooperation with ISVs and integrators, develop highly targeted industry solutions and professional application software. management software and security software vendors should also see such business opportunities to accelerate support for Linux and cooperate with Linux system vendors.

Another prominent problem is the support of technologies and services. In a telephone interview, many users thought that the services provided by Linux system vendors were not in place. Although they were open-source systems, once they were purchased as commercial systems, adequate technical support and services should be provided. Users generally want to provide on-site services to solve problems they encounter in their applications, some senior technical engineers are also required for technical consultation or training.

Some users also mentioned security issues. They believe that Linux does not have as many users as Windows and can promptly remedy any security problems they find. That is to say, linux still lacks some security mechanisms and timely security remediation measures to prevent some types of attacks. Therefore, Linux urgently needs more security software support.

Figure: 36p18c.jpg

For a long time, the cost of Linux has always been the most powerful weapon for many supporters to promote it to enterprise-level core applications. system, database, and middleware vendors all believe that, linux is a free replacement product that replaces old operating systems such as Windows and proprietary Unix, especially after the combination of Linux and IA-based hardware systems, its cost advantage is even more prominent. But the fact is that everything is priced, including open-source software. Moreover, any middleware or database built on Linux won't be cheap because it is Linux. Here we only look at the Linux itself. Is it really cheap to calculate the purchase, migration, operation and support fees? In other words, does Linux have a lower TCO than Unix or Windows?

Part II

Linux fees

The simple answer is as follows: the more comprehensive an enterprise uses Linux on its IT infrastructure, the more economic IT may gain from its early investment in such an operating system. These investments include the cost of Linux training and tools and migration from Unix or Windows environments, the economic benefits will increase with the emergence of better management tools, more third-party vendor support, and more skilled Linux system administrators.

The benefits of migrating from Unix or Windows to Linux are driven by four main types of fees: purchase, migration, management, and support. The following briefly discusses some key issues in various expenses.

Hardware and software fees

One of the main factors driving the adoption of Linux is the cost savings produced by replacing the expensive Unix System Based on the RISC processor with the IA Architecture Machine of large-scale production. Computing hardware costs are easy, says Ted Schadler, chief analyst at Forrester Research. "Thanks to IA's economics, transferring Unix workloads to Linux is an easy decision, the price difference between the hardware system of the RISC architecture and the hardware system of the IA architecture is 5000 ~ Between $25000.

In terms of software, the cost difference is not obvious. A study by IDC, Meta Group, and Robert Frances Group found that the Linux license fee is lower than the Windows License fee. However, some analysts believe that this is not a comprehensive one-to-one comparison. Schadler of Forrester pointed out: "when developing an application, it is not a problem to choose Windows or Linux. This is the decision to choose Java on Linux or Windows. Microsoft to Windows, SQL Server, and. net Framework is bound with a lot of things. If you plan to develop a common application and deploy it at any price, Windows will not struggle to win, because it will get so many bundled items."

In terms of hardware and software, the cost advantage that Linux is often overlooked is the flexibility it provides on the road to future migration and upgrade, because when using Linux, you can master your own upgrade cycle.

Migration fee

When considering migration from Unix or Windows to Linux, enterprises should carefully analyze the one-time migration cost. One of the biggest costs is to train system administrators to master Linux as soon as possible.

In fact, Unix skills are very similar to Linux skills, which reduces the cost of migrating from Unix to Linux. For a Windows-only user without any Linux skills, there are huge obstacles. Those who have Unix skills but cannot adapt immediately can easily download Linux and try it at home.

Because the minimum re-training cost and the hardware procurement cost are greatly reduced, migration from Unix to Linux is generally a reasonable economic choice. However, due to the high cost of re-training and the low cost of hardware purchase, migration from Windows to Linux is more dependent on luck.

Other migration costs include code that may have to be rewritten to reproduce the existing functions on the platform that Linux intends to replace, data that must be migrated, work on integration with the backend system, and software that must be purchased; depreciation is a hidden migration cost that many people don't think of. Some middleware, such as application servers, can be easily transplanted to Linux. If the original application already runs on such middleware, the one-time migration cost will be reduced.

Management Fee

The maximum cost of most Linux TCO studies so far is the personnel cost required for future operation system management. In a comparative study of Windows and Linux conducted by IDC in 2002, staff costs accounted for 62% of the total cost of these two environments in 5 years, which also caused the maximum cost difference between the two systems, windows takes the lead at a lower cost.

Although most analysts believe that the cost of Windows administrators is lower than that of Linux administrators, the actual debate focuses on the cost benefits of integration and the availability and quality of management software tools for Linux platforms.

Oracle's point of view is that users who simply use Linux with other operating systems may not save money, and puts forward the points of integration: if you fully migrate to Linux and purchase or develop robust management tools, you can save more costs on Linux, because Linux developers can use the Linux operating system without restrictions to manage more machines and larger workloads, they only need fewer administrators.

Some users of large Unix systems believe that on Linux, tools, processes, and automation functions have not yet developed to the level they have reached on Unix, so the cost for migrating to Linux is zero.

Duncan, chief technology officer of Cedars-Sinai at a medical center, found that some key cost-saving tools were missing, although he was satisfied with the overall migration of his institution to Linux. "Backup has always been an annoying problem," Duncan said ." He pointed out that, as part of IBM's Tivoli Storage Manager, the Medical Center has been using Hot Backup Agents on Oracle databases running on AIX and Windows NT, but such agents cannot be used in Linux. Duncan also found that the server may have problems with SAN support.

The cost of security management seems to have always been a better management expense field than that of Windows. Duncan said that an important reason for Cedars-Sinai's migration to Linux is due to hot Patching, service packages, and other things. "We need to make a lot of changes on the NT server, the time we spent modifying the NT server was really scary. We had to be able to set the server and ignore it. From this point of view, we migrated to Linux."

Support fee

Linux supporters said that the cost of Linux support is lower, and more vendors provide support for it. They said that Linux machines often run for several years without a restart.

When a new Linux machine crashes, either change the machine because the cost of the change is cheaper than the repair cost, or use a method called "Google service, find a solution on the Web newsgroup and message board. Enterprise Users are gradually eliminating their initial doubts about this support method. In the past, Linux support was a hindrance to enterprise adoption, and is now becoming a strength.

Users may choose from hardware or software vendors, distribution vendors, or third-party support, which may occur on Windows. However, when using Microsoft Solutions, providers cannot actually modify Windows for you, all Linux support vendors have the same capability.

Table: 36p19.doc

Linux has been widely deployed in many data centers and is usually used as a Web server or file server. It can also be used to process network tasks such as DNS and DHCP, but as a platform for running key enterprise applications, Linux does not seem to be favored. It is usually used for Solaris, AIX or HP/UX, provides the highest level of performance and scalability for applications. However, many people think that with the release of the new Linux v2.6 kernel, this situation will change. That is to say, the new Linux v2.6 kernel will become a powerful boost for Linux to enter key enterprise-level applications. So what exactly does the new v2.6 kernel bring?

Next

Larger and stronger kernel Linux v2.6

The Linux v2.6 kernel has created a new era that supports large computing environments and higher loads, enabling Linux to perform difficult tasks currently executed by Solaris, AIX, or HP/UX.

The main feature of v2.6 kernel is to support large-scale server architecture. It supports a maximum of 64 GB of memory and can process file systems larger than 2 TB, and supports 64 CPUs in x86 SMP systems, these features make the kernel and Linux more qualified as the running platform of the critical task system. The new kernel also supports the NUMA (non-uniform memory access) system, the next-generation SMP architecture, and PAE (physical address extension ). This kernel breaks some manual restrictions imposed on Linux from the very beginning. It supports up to 4096 primary devices, and each primary device has up to 1 million secondary device numbers, for most users, these numbers far exceed the actual requirements, but for enterprise systems that need to deal with many devices, this solution is a huge leap. In addition, the new v2.6 kernel also contains NPTL (local POSIX thread Library), which is used to replace v2.4's LinuxThreads. NPTL enables Linux to support enterprise-level threads, its performance far exceeds the level of LinuxThreads.

Born for speed

Before the v2.6 kernel is released, Linux runs the task on a first-come, first-served basis, which interferes with the kernel's interrupted task to process other processes or functions. However, the v2.6 kernel can take priority when needed, allocate resources for jobs that need immediate processing, and then continue previously interrupted tasks. This type of interruption only happens within a few minutes, and is usually not noticeable, but it makes people feel that the overall system performance is very smooth. The v2.6 kernel does not make Linux a real-time operating system, but it is constantly working to ensure that tasks are processed and completed at any time as needed.

The core of these enhancements is a new process scheduler. The process scheduler in the kernel can allocate CPU resources between system processes. The performance of the scheduler directly affects system response and process latency. In the v2.6 kernel, the new 0 (1) scheduler incorporates new algorithms to improve system performance, especially the performance of interactive tasks. V2.6 kernel has two new I/O scheduling programs. By default, a predictive program is used to effectively improve the performance of I/O scheduling, ensuring that all processes can obtain the required I/O time without queuing. The other is the time-limit scheduler, which can allocate time to requests in three queue modes. The predictive scheduler will only attempt to predict the process I/O requests before these process requests are actually sent. No matter which method is used, the new process and I/O scheduler have greatly exceeded the level of v2.4.

Figure: 36p0000jpg

V2.6 kernel Test

In actual situations, it is very helpful for you to fully understand the new kernel in the environment where the user is most likely to apply it to test the new internal check.

Network World has done a comprehensive test to compare the performance of the 2.4 and 2.6 kernels. The test runs on three completely different hardware platforms: Intel Xeon (x86), Intel anteng (IA-64), and AMD Opteron (x86_64 ). The basic operating system uses Red Hat Linux Enterprise Server v3.0, but the kernel test requires a customized kernel compiled for each Server. V2.4 uses the official v2.4.23 kernel, while v2.6 uses the official v2.6.0 kernel. To run on different platforms, the kernel must be compiled in a targeted manner, such as using the x86_64 patch for the AMD64 platform for x86-64.org.

The file sharing test is designed to simulate standard Samba server loads, based on Samba v3.0.1 with local verification. In the Xeon system, the average throughput of the v2.4 kernel is 38.85 MBps, while that of the v2.6 kernel reaches 67.30 MBps, an increase of 73%. The anteng test also showed similar performance differences between the two kernels, with v2.6 leading by 52%. On the Opteron system, the kernel speed of v2.4 is 49.37 MBps, and the result of v2.6 is as high as 72.92 MBps, increasing by 48%.

The performance improvements in Samba testing are largely attributed to the improved scheduler and I/O subsystem in the v2.6 kernel.

The database test environment is based on MySQL v3.23.58 and runs on the SQL Benchmark Test Set provided by MySQL. As a whole, the performance of v2.6 kernel has completely exceeded that of v2.4 kernel in database testing, especially on the Apsara stack server. Compared with v2.4 kernel, the performance of the new kernel version has increased by 23% (519 s ahead ). On the Xeon platform, v2.6 showed a 13% improvement (200 s ahead ). On the Opteron server, the speed has increased by 29% (415 s ahead ). Among all tests, the most impressive is table insertion. The v2.6 kernel improves performance on Xeon servers by up to 10% (100 s ahead ), the performance on the Opteron and anteng platforms is extremely powerful.

The Web server test also showed great improvements. Static Page testing uses an HTML page with a size of 21.5KB, containing two 25KB images served by Apache 2.0.48. The purpose of the test is to use the Apache AB benchmark tool to measure the number of requests per second. On the Xeon platform, the v2.6 kernel is less than 1000 requests/second ahead of v.2.4, with an increase of 40%. On the CERT server, the performance of v2.6 has increased by 47%.

Web application testing uses a custom CGI script written in Perl, pointing to a MySQL database running on the same system. The overall figure shows that the result is much smaller than the static test, but the Opteron test is an exception. The general performance improvement of all platforms is between 14% and 22%.

Development Direction

Experts pointed out that IT enterprises running large databases and other key task applications on the v2.4 kernel should begin preliminary tests on v2.6. V2.6 kernel is the future leader. Enterprises should get familiar with the capabilities of this new kernel and make some preparations for future applications.

To make full use of the enhanced features of v2.6 kernel, the administrator needs to configure and establish a customized kernel for the server architecture. However, this type of kernel has the following problems, they differ from the officially released kernel, regardless of the default options or included patches. On the other hand, these kernels usually support a wide range of hardware because almost every module is configured to ensure the hardware compatibility of the target system.

At the same time, v2.6 kernel also has many defects, and kernel developers are actively looking for and solving these problems. If you begin to consider porting to the v2.6 kernel, we strongly recommend that you perform a comprehensive test of the new kernel before any productive implementation.

The v2.6 kernel is proud to have many new features and significantly improves performance. This kernel will also make Linux ready to meet the challenges of the new phase: real enterprise-level key applications. In order to seize the data center market, Linux must be able to adapt to changes in the needs of existing users at any time, while constantly expanding new areas and bringing more users to the Linux camp. From the current situation, the v2.6 kernel is ready for this task. -

Editing and editing ideas

Calcium Supplementation required for Linux

-Jiang Bo

There are indications that Linux has grown up. In particular, from last year to 2003, it was the next year of Linux, and it began to become popular, rather than "always developing and growing", as previously said ". A recent Linux research report by Forrester pointed out that the technical barriers hindering the popularization of Linux are disappearing. It already has 4-chip support, enterprise-level thread mode, and security capabilities, in addition, Linuxcare, Opsware and other vendors have also developed new data center-level management tools. These factors make Linux a turning point in popularity.

However, like any other things, the growth of Linux is not smooth, and its ups and downs are manifested in the problem of "calcium deficiency", specifically its lack of support and management tools.

We also mentioned these two problems in the text. I would like to reiterate here that we hope that vendors that support and strongly promote Linux can quickly solve the problem. The Forrester survey report also pointed out that 46% of respondents believed that Linux's support was insufficient, the support and service quality obtained from independent Linux vendors cannot be compared with the support services of related hardware systems and solutions provided by industry giants such as HP and IBM. Of course, everyone has been working hard. For example, Oracle promises to "single point of support" for users ", red Hat, Novell SuSE Linux, and Red Flag Linux, which support Oracle database products, provide comprehensive support, that is, as long as a user calls in, oracle provides services for both database and operating system problems. However, at present, the support problem is still a big problem. Independent Linux vendors need to redouble their efforts in this regard, mainly to provide localized support and services and industry solutions, including expert-level on-site support, some popular training, specialized and industry training, etc.

There is also a lack of calcium administration tools. Forrester's research report warned that lack of management is a growing headache for Linux. Various Linux software packages, even those sold by the market leader, cannot be exchanged. In addition, tools for backup and security management are lacking.

From: blog China

Related Article

Contact Us

The content source of this page is from Internet, which doesn't represent Alibaba Cloud's opinion; products and services mentioned on that page don't have any relationship with Alibaba Cloud. If the content of the page makes you feel confusing, please write us an email, we will handle the problem within 5 days after receiving your email.

If you find any instances of plagiarism from the community, please send an email to: info-contact@alibabacloud.com and provide relevant evidence. A staff member will contact you within 5 working days.

A Free Trial That Lets You Build Big!

Start building with 50+ products and up to 12 months usage for Elastic Compute Service

  • Sales Support

    1 on 1 presale consultation

  • After-Sales Support

    24/7 Technical Support 6 Free Tickets per Quarter Faster Response

  • Alibaba Cloud offers highly flexible support services tailored to meet your exact needs.