Woman checking server rack size

8 reasons for data center to consider adopting super converged infrastructure

Compared with the server configuration method of Japan, super fusion has advantages in speed and simplicity.As enterprises move their workloads to the cloud, the demand for local data center equipment gradually decreases. However, local deployment is far from disappearing, and Hyperconverged Infrastructure (HCI) is one of the booming areas. 2020-09-16 10:26:32

Compared with the server configuration method of Japan, super fusion has advantages in speed and simplicity.

As enterprises move their workloads to the cloud, the demand for local data center equipment gradually decreases. However, local deployment is far from disappearing, and Hyperconverged Infrastructure (HCI) is one of the booming areas.

HCI is a scalable, software integrated infrastructure that uses a modular approach in computing network and storage capacity. HQI does not use dedicated hardware independently, but uses distributed horizontal commercial hardware and provides a single panel dashboard for reporting and management. It has a variety of forms: enterprises can choose to deploy hardware independent hyper converged software provided by Nutan "x" and VMware, and can also deploy integrated HQI devices provided by HP, Dell, Cisco, Lenovo and other suppliers.

Its market is growing rapidly. By 2023, Gartner expects that 70% of enterprises will run some form of hyper converged infrastructure, compared with less than 30% in 2019. As HCI becomes more and more popular, cloud providers such as Amazon, Google and Microsoft provide connection services with local HaI products to realize hybrid department and voice management.

Why is HCI so popular? Here are some of the main reasons.

1) Simplified design

The traditional data center design includes discrete storage and independent server layers, as well as a dedicated network connecting these computing and storage islands. This was the case before the cloud age, but in the cloud age, it was too heavy. Greg Smith, Nutan ix's vice president in charge of product marketing, said, "It takes several weeks or even months for Department T to provide new infrastructure, and then the development department will develop new applications and launch them to the market as soon as possible, but this is not feasible. HCI has fundamentally simplified the architecture and operation and maintenance of the data center, reducing the time and cost of managing and delivering applications.

2) Cloud integration

Nutan "HCI software provided by companies such as VMware is deployed in the customer's data center and cloud instance in the same way;"; It runs on the cloud's raw server instance in exactly the same way as it does in the data center. Smith said, "If enterprises want to build a hybrid cloud, HQI is the best foundation. They can deploy applications in their own data centers and integrate them into the public cloud."

He added: "Because both sides are the same, I can let a department manage the end-to-end hybrid cloud, and I believe that any application running in my private cloud can also run in the public cloud environment."

3) Be able to grow from small to large

Tom Lockhart, the II system manager of Hastings Prince Edward Public Health Company in Belleville, Ontario, Canada, said, "The vegetable HQI can integrate computing, network and storage into a box, and quickly and easily expand this solution without a lot of downtime."

In traditional methods, multiple hardware component servers, Fibre Channel switches, host based adapters, and virtual machine managers must be installed and configured separately. With super fusion technology, everything is defined by software. HCI uses the storage in the server. The software is almost completely automatically configured, and automatically detects the hardware to establish connections between computing, storage, and the network.

Chad Dunn, Vice President of HI Product Management, said, "When we work hard, customers usually have a very good experience. A few months later, they tested several workloads, and then began to expand them from the data center to remote sites.

Dunn added: "At the beginning, they can be smaller and expand gradually, but their O&M mode experience is unique - no matter whether they have 1000 nodes or 1000 sites, each site has 3 nodes, no matter whether they have 40TB data or 40B data. Their software updates can be consistent. Because they use the same tool set, there is no need to train employees.

4) Scale reduction

Starting small, customers found that they could reduce the hardware stack to the amount they needed without providing too much capacity. The elimination of mutually independent methods also allows users to cancel some hardware.

Josh Gooda, automation engineer of the steel manufacturer USS-P0SCO Industrial Company, said that his company's deployment of HI was mainly due to the ability to expand clusters, that is, hardware clusters can be located in different physical locations but connected together. This is primarily for backup, so if one site fails, another can take over the workload. In this process, USS pSCo removes many expensive hardware and software. Goodall said, "We have cancelled several CpU software licenses, cancelled SANs at other sites, and do not need SRM (Site Recovery Management) software or Commvault licenses. Our annual license renewal saves us 25 to 3 dollars.

5) No special skills required

To run a traditional three-tier environment, enterprises need experts in computing, storage, and networking. With HCI, ordinary technical consultants and employees can manage the environment, and enterprises do not need to hire experts with high salaries.

Smith said, "Ha makes us versatile. We don't need to hire storage experts and network experts anymore. We must all have infrastructure, which makes the actual maintenance of infrastructure much easier than in typical scenarios. In typical scenarios, we need deep professional knowledge to do well in sub management."

Lockhart of Hastings Prince Edward Public Health said that it is also much faster to add new computing/storage/network devices than traditional infrastructure. He said, "It takes 20 minutes to upgrade to our server cluster, and there will be no downtime. However, the traditional method requires downtime and service interruption for several hours.

Lockhart added, "You don't have to focus on infrastructure. You can spend more time and resources on the workload, so that the business can be more valuable. When you don't have to worry about infrastructure, you can have more time to add value for customers."

6) Shorter disaster recovery time

The key elements of super fusion products are their functions of backup, recovery, data protection and duplicate data deletion, as well as the analysis function of comprehensive inspection. The disaster recovery component is managed by a single dashboard. HCI monitors not only local storage, but also cloud storage resources. With data deduplication, the compression ratio is up to 55:1, and the backup can be completed in minutes.

Uss pSCO Industries is an HP enterprise store that uses HPE's SimpliVictory HCI software, including deduplication, distribution and recovery functions. Goodall said that the mixed workload was compressed by about 12~15:1, so it was no longer necessary to use the second party backup software.

More importantly, recovery times have also been reduced. Goodall said, "The best example recently is that a Windows update messed up a security policy. We haven't found this error for several weeks. In about 30 minutes, I completed the original four week backup, updated the system, restarted and tested a 350GB system. Previously, it was a process of several hours just to restore a backup.

7) Hyperfusion analysis

HC products come with a lot of analysis software to monitor workloads and find resource bottlenecks. Monitoring software is integrated into a single dashboard view of system performance, including negatively impacted performance.

Hastings recently had a problem with Windows 7 migration, and it is easy to obtain performance information using the H model. Lockhart said, "The problem is that, depending on the time of the day, the workload running leads to insufficient memory and excessive CPU queuing and paging. We can figure out the whole problem in an hour. It is easy to determine the problem. If there is no single view, it may take longer."

8) Reduced time to manage network and storage resources

Goodall said that he used to spend as much as 50% of his time dealing with storage problems and backup matrices. Now it only takes about 20% of the time, and most of the time is used to deal with and solve the problems of the old system. And his application performs better under HCI. "There is no problem with our SQL database. If there is, the performance has improved a lot. This is due to the migration to an all solid state disk (not a mechanical disk) and the data deduplication function, as well as the reduction of read and write operations in the environment."

Andy Patrizio is a freelance writer in Southern California. He has been engaged in the computer industry for more than 20 years. Every x86 personal computer he owns is his own assembly, not including notebook computers.