4 reasons to risk moving to software-defined data center – y

Over the past 3 years, only the EMC federation has helped deploy software-defined data centers in more than 1,000 organizations around the world, fully preserving the existing infrastructure, optimizing its use to support business processes. And according to the observations made, in most cases, companies come to such a decision quickly and even unexpectedly for themselves. Therefore, even if now the software-defined data center (hereinafter, for brevity, POTSOD) seems to you to be something too outlandish, do not rush to swear by its implementation. The risks of making major changes in the IT organization are intuitive. This article discusses the benefits of moving to PODSOD and 4 common reasons for building them in organizations.

What are program-defined data centers, in general, everyone knows. And, although there may be discrepancies in the nuances, each respected IT manufacturer orients the strategy of its portfolio of solutions in order to help its customers transform “ordinary” data centers into software – defined ones. In this case, first of all, we are talking about the economic benefit of customers from such a transition. However, in the process of justifying before the business the necessity of this transformation with an enviable regularity, difficulties arise. This is not surprising, given the complexity and intricacy of the interconnection of one world with another. I remember a bearded anecdote about how my son asked the programmer’s father why the sun rises in the east and sets in the west. The answer “if it works - do not touch anything”, characterizes reality well. The sad and ridiculous thing is that corporate IT, once the vanguard of innovation, is now more associated with the shamanic caste performing rituals with tambourines.


')
And, nevertheless, the need for change has matured. It manifests itself primarily in the fact that traditional business, in a broad sense, is repeatedly exposed to attacks by young and daring startups, drawing on energy, opportunities, and even business ideas from new IT areas. There is no need to go far for examples. Look at the apps in your smartphones. You will see few “ordinary companies” there, in which many of us, by the way, work.

Adult players either try to vigorously adapt to new realities and remain competitive, or more often they go forward by inertia while it is there. The tragedy of the IT departments is that, having a powerful arsenal of tools to build a new model of technical support for business, they often do not have the appropriate apparatus to initiate this process in the company. And business, in some places, is already seeing IT specialists as lovers of expensive toys. Although they are being thrown money from time to time, they have not been trying to figure out for a long time what pipe they fly into year after year. And when it comes to major changes not only in architecture, but also in business processes, there are almost no chances to come to an agreement.

It is no secret that serious changes in infrastructure or business processes often cause some kind of critical failure, or even a series of failures. And this can have unpleasant long-lasting consequences in the presence of a high load on the infrastructure.

Therefore, in spite of the fact that for many specialists, software-defined data centers have, at least in theory, a number of advantages, in practice, few decide to deploy them. In addition, you can not just take and replace the components of the existing infrastructure, starting with a clean slate. The transition process requires careful thought and design, taking into account existing processes and architecture. Therefore, if these changes begin to push, there is a natural resistance.

In Russia, this is even more difficult, with our "special way" and "Russian specificity", which is not understood "in the West." The fact is that we, seeing ourselves from the inside and looking at Western IT – models through their advertising booklets and press releases, conclude that we are not just very different with them, but they are aliens in general. And all that is good for them, for us, for earthlings is deep space. I was of the same opinion, until I realized that although there is a gap between "our" and "their" IT, it is far from cosmically large. So, perhaps, we have something to take for ourselves from their advanced developments, subjecting, if necessary, to critical thinking. Further, I propose to consider and analyze four arguments in favor of the transition to POTSOD – y.

But first, I want to remind you of the features that modern POCS implementations have:

• Automated deployment of virtual machines, network resources, storage resources . Increases the speed of deployment and provision of new services. It is possible to use global policies to manage workflows, including, for example, setting up backups, security policies, network settings. Getting rid of a lot of other routine.
• Automated storage tier management . High speed of work with data is provided, while keeping storage costs at an acceptable level. The ability to use SLO policies for storage management.
• Automated monitoring . Providing information about the status of services in real time. Facilitates the management POCOD. Thanks to responsiveness and the possibility of proactive monitoring, increases reliability, availability and performance. Due to the increased efficiency of the service, operating costs are reduced.
• Transparency of resource use . The ability to accurately assess the cost of supporting services, the correctness of the calculation of resource consumption by various departments of the organization. Reducing the risk of unjustified excessive consumption.
• Automated crash protection, including auto repair . Preventing service failure, or accelerating its return to service. Reduces the risk of operator error during recovery.
• Built-in backup and recovery services . The ability to use personal data protection policies, including in accordance with corporate or foreign policies. Deduplication during backup. Minimize costs by accelerating backup and recovery.

Thus, PODSOD is an interesting technical solution that significantly simplifies life and simplifies the performance of various tasks, potentially allowing for a new level of IT-services for businesses. However, both "we have" and "they", real hardcore experts do not get tired of repeating that they can do "all the same with their own hands, for their salary." It has its own truth. Moreover, if you come to the managers with an attempt to justify such a decision, then at best we will encounter sincere misunderstanding. Like, "before you demanded money to do your work, and now you are asking for money in order not to do the work."



It turns out that the situation for the transition to PODSOD should mature, at least, to the state when the business "cannot live the old way." And ideally, when it still does not want IT. And this trend is gradually evolving.

Over the years, working with all sorts of customers around the world has accumulated some statistics on the circumstances that still push companies to deploy POCS. Let's take a closer look at them and try to try on ourselves.

Reason number 1. Performance problems combined with inefficient use of infrastructure

Of all the reasons, this is perhaps the most common, including in Russia. Because it is rare to find an IT manager who actually knows exactly which applications and how many resources they consume, and in detail. The situation is widespread when resources are allocated for critical applications based on their peak needs, plus some margin. As a result, the infrastructure is overloaded with iron, which, on the one hand, is idle most of the time, and on the other hand, it is experiencing regular bursts of stress, which is stuck in narrow places, which creates a feeling of lack of physical resources. And they just need to properly plan and use. The desire to overcome this situation leads to the consideration of program-defined solutions, because the lack of information does not allow to form an infrastructure development strategy, plan a budget, etc.

Reason number 2. It is difficult to determine and maintain compliance with the service level of the application and the physical resources provided to it

Sooner or later, business has to change following the demands of the market. This is reflected in new projects, application development and in the use of new data in business processes. As a result, workloads change, data volumes grow. It happens? Yes!

In order for organizational processes and infrastructure to adjust to dynamically changing workloads, it is necessary to assess in a timely manner how to ensure maximum efficiency in the use of resources. In a typical POTSOD architecture, special software is used for this purpose, which automates or at least facilitates this process.

Some solutions also allow you to manage, interact and automate resources , even if they are purchased from different suppliers.

Once there was a case when one of the big customers found a number of applications working with Tier 1 storage, which, after evaluating the load profile, were transferred to a lower level storage without any loss of performance. This brought a savings of $ 5.2 million to the customer. I don’t think that this could have been done easily if the IT infrastructure of this customer was rigidly tied to hardware levels. But using POTSOD - easily. In Russian practice, we also had such a similar case. Only as a result, we were not amused, because the customer happily reduced the volume of purchases of our iron. True, he was very grateful to us.

Reason number 3. IT employees are overworked, recruitment is suspended, and demands are growing

Most often, data center management is weakly automated. Sometimes it all comes down to sending alerts about the achievement of a certain level of resource use by a particular node or application. If there is an extensive infrastructure on which many applications are deployed with different requirements and restrictions on the use of resources, the analysis of all these warnings requires a lot of time, human attention and correct response.

One of the most important advantages of POCS in solving this problem is the use of global policies and the gradual transition to automated allocation of resources in accordance with them. By itself, the classification of applications is a very useful measure, allowing to determine the blind spots of the infrastructure. And when it will be possible to allocate resources on the basis of the received classification and measure the success of reaching service levels, the business opinion on IT will improve.

On the other hand, it becomes possible to use the released people to solve more useful and important tasks, providing them with a tool to manage much larger amounts of resources. For example, according to Gartner, on average, one employee handles data about 250 TB. According to the results of cooperation with a number of major service providers implementing the PEDSOD concept, this figure rises to 600 TB, and in some cases to 900. In the future, thanks to even deeper automation, the 1 PB mark will be passed.

Reason number 4. "Shadow IT"

Today the use of public clouds by enterprises is growing. Yes, yes, in Russia too. For example, to provide development flexibility. Often, this is even tacitly welcomed, because many developers do not believe that they are able to effectively and profitably use their own IT services. For them, the company's internal IT systems are expensive, long and, sometimes, of poor quality. Therefore, they use public IT-resources at their discretion, and, most unpleasantly, without the knowledge and control of their own IT services. This can lead, at a minimum, to non-observance of service levels, and as a maximum, to loss of intellectual property, leakage of personal data, etc. Nevertheless, corporate IT is easier to head this outrage than to try to prevent it. And for this you need to prepare your infrastructure for secure integration with public clouds, or learn how to provide IT services to internal users with the quality and convenience that they receive from public providers.

For example, to minimize the risk of uncontrolled price growth for external services or the lack of stability of their work, assess the potential for deploying its own cloud storage . According to our experience, the use of software – defined storages is not only not inferior in terms of flexibility to external services, but in some cases costs up to 25% cheaper. In addition, the service infrastructure is completely under your control.

Finally

The transition to a new form of IT organization is mainly due to the following factors:

1. The need to continuously solve performance problems, combined with the desire to more effectively use the IT infrastructure.
2. The requirement to maintain compliance with the application service levels and the physical resources provided to them.
3. The desire to more effectively use the existing IT human resources in a situation of constant pressure "from above".
4. The growing risks of "shadow IT".

PODCOD is not just a data center consisting of virtualized hardware. It is the automation of routine operations, reducing the risk of human error and providing service levels. This is an effective use of existing infrastructure and a more structured understanding of IT costs.

Of course, not always such a transformation can occur solely on the initiative of the IT department. Without understanding and support at the business level, such events are hardly possible. But that's another story.

Used sources:
1) https://www.linkedin.com/pulse/four– catalysts – software – defined – data – centre – rollouts – dayne– turbitt?published= t
2) https://www.linkedin.com/pulse/software–san–driving–commodity–storage–enterprise–antonio–rome?published=t
3) http://www.emc.com/collateral/analyst–reports/idc–sds–3rd–platform.pdf
4) http://www.emcfederation.com/

Follow our blog on Habré and come to visit me!
Denis Serov