Entry
2019 is slowly but surely coming to its end. The IT industry continues to develop actively, delighting us with a large number of new technologies and along the way replenishing our vocabulary with new definitions: Big Data, AI, Machine Learning (ML), IoT, 5G, etc. This year, Site Reliability Engineering was discussed especially often. (SRE), DevOps, microservices and cloud computing.
Some technologies, such as Blockchain and cryptocurrencies (Bitcoin, Ethereum, etc.), seem to have already passed the peak of their popularity (hype), so the general public has been able to take a more sober look at them, identifying their positive and negative sides, as well as deciding where and how best to use them. A balanced look at the topic of Blockchain and cryptocurrencies can be found in. I highly recommend checking it out.
Other technologies are only gaining their popularity so far, forming active communities around them, including not only supporters and adherents, but also ardent opponents.
Is everyone going to DevOps?
A special mention from me today will go to DevOps, a new approach to software development and operation. There were indeed many articles and disputes on this topic this year.
The term DevOps today is interpreted quite broadly. Someone under DevOps understands a special approach to the development and operation of software, when people who can both code a little and admin are involved in the work. For others, this is, first of all, the presence of a personal system administrator in the team, which allows you to remove part of the non-core load from software developers in the form of setting up the system environment, creating test environments, implementing integration with internal and external services, as well as writing automation scripts. For others, it's just a set of trendy technologies and tools that you need to use to stay always young and successful. For the fourth, it is CICD and everything connected with it. DevOps really has a lot of interpretations, so anyone can independently find in them what is closer to their liking.
Different interpretations of DevOps generate heated discussions, which leads to the appearance of more articles on this topic. Some of them I even saved to my bookmarks:
- .
- .
If you read enough articles praising DevOps, you might get the impression that any system administrator/engineer needs only to change their current position on their LinkedIN profile from DevOps engineer administrator, and invitations to interviews from HR from large and successful companies will immediately begin to pour in. , who will promise a salary 2 times higher than the current one, will give out a brand new Macbook, a gyro scooter and will not forget about a subscription to free vape refills and an infinite number of smoothies. In general, IT paradise will come.
If you read articles that belittle the merits of DevOps, then another impression begins to take shape that DevOps is a new kind of slavery, where people have to code almost on the same level as developers, help them fix bugs, do automation and CICD, deploy Jira from Wiki , spin clouds, collect containers and manage them, while doing admin work in parallel, not forgetting about refilling cartridges, crimping twisted pair and watering office flowers.
But, as you know, the truth is usually somewhere in the middle, so today we will try to understand a little about everything.
No more admins?
As a system administrator/engineer who has been working with Microsoft and VMware products for quite some time, I began to notice that in the last few years there have been talks periodically that no one will need system administrators soon, because:
- The whole infrastructure is about to change and turn into IaaC (Infrastructure as a code). Now there will be no GUI with buttons, but only PowerShell, yaml files, configs, and so on. If a service or its component breaks down, then it is no longer necessary to fix it, because it's faster to deploy a new copy of it from the last working state.
- The entire IT infrastructure will soon move to the clouds, and locally (on-premise) there will be only network cables to the nearest router that will connect us to all other corporate resources located in the cloud. Well, at most the printer will remain locally so that the girls from the accounting department can print images of cats from the Internet on it. Everything else should be in the cloud.
- DevOps gurus will come and automate everything around, so admins will only remember with warmth in their souls how in the old days they ran pings and traces to diagnose basic problems on the network and on servers.
- I also heard about such a phenomenon as "Vendekapets", but it was quite a long time ago, at the dawn of my career, when I was just starting to take my first steps towards system administration. But for some reason, "Vendekapets" never came, like the end of the world according to the Mayan calendar. Coincidence? Don't think. π
Is it possible that Windows system administrators, who work closely with Microsoft products today, will soon be of no use to anyone? Or is there still a need for them? Will Windows administrators continue to wear their status of administrator-engineers, or will they be relegated to the role of a low-skilled workforce a la anykey (give, give, fetch)?
Even here on habr.com in the System Administration hub, we only see mentions of kubernetes, linux, devops, docker, open source, zabbix. Where are the words beloved to our hearts Windows, Active Directory, Exchange, System Center, Terminal, Print Servers, File Servers, bat and vbs scripts, or at least powershell. Where is all this?
So is there life after Windows or Windows system administrators should quit everything now to learn Linux, docker, kubernetes, ansible, python and go to DevOps?
Maybe everything is fine with Windows, just now there is a temporary hype of the Linux + docker + kubernetes + ansible + python bundle, which overshadowed our beloved Windows? What does a Windows system administrator need to do in 2020 to be in demand in the labor market?
Unfortunately, there are more questions than answers here, so the current article will try to help us sort things out a bit. The article is primarily devoted to Windows administrators and engineers, but I am sure it will be of interest to other IT specialists as well.
Microsoft goes to the clouds?
The Windows admin is, first of all, an adherent of Microsoft, so further we will talk about it and its wonderful products.
Microsoft has a fairly wide portfolio of software solutions, many of which are leaders in their niches. If you work as a Windows administrator as an engineer, then most likely you have come across them one way or another. Below I will give a brief description of each of the products and describe the possible prospects for their development for the next 3-5 years. This is not a secret insider from the headquarters in Redmond, but my personal opinion, so alternative points of view in the comments are welcome in every possible way.
Local installations (on-premises)
Microsoft Exchange Server is a multifunctional mail server that includes not only work with mail, but also with contacts, calendars, tasks and much more. Exchange Server is one of the flagship products of Microsoft, which has become the de facto corporate standard in many companies. It has tight integration not only with the products of Microsoft itself, but also with solutions from third-party vendors. Exchange is popular both in medium (from 100 people) and large companies.
At this point in time, Exchange Server 2019 is considered the current version. Previously, the product developed quite actively, but starting from the version of Exchange 2013, this development has slowed down a lot, so Exchange 2016 can be conditionally called Service Pack 1 (SP1) for Exchange 2013, and Exchange 2019 - hence Service Pack 2 (SP2) for Exchange 2013. The fate of the next on-premise version (Exchange 2022) is still in question.
Now Microsoft is actively promoting Exchange Online as part of the Office 365 cloud service, so all new features appear there first of all. Exchange Online is not only the first to get new features, but also acquires additional features that will not be transferred to local installations (on-premise) in the near future. This is done in order to speed up the transition of a number of companies to the clouds, because. The subscription model is more financially beneficial to Microsoft than a one-time sale.
If you are currently maintaining a local installation of Exchange Server (2013 - 2019), then you can continue to do so for the next 3-5 years. Along the way, it's worth starting to explore the possibilities that Exchange Online provides; and hybrid configurations, this is when the local and cloud versions exist at the same time. Even if we assume that the next on-premise version of Exchange will no longer exist, then the knowledge gained now on Exchange Server will continue to be relevant for some time to come for a number of reasons:
- The number of local installations is currently quite large, so qualified administrators will be needed to support them. Not all organizations will be able to move their mail to the clouds in the near future for one reason or another.
- Cloud migration projects are not yet trivial, so knowledge of the specifics of both on-premises and cloud solutions is required to bypass most of the pitfalls and successfully complete the migration.
- Knowledge of smtpimapmapipop3, mail flow, dkim, dmark, spf, antivirus, antispam protocols is universal and will be applicable to any mail systems.
- The experience gained while working with an on-premises Exchange Server will make it much faster to understand Exchange Online and set up the desired configuration.
- E-mail is one of the most important channels of communication with the outside world, so the need for it will remain. Adepts "messengers and chat bots will replace mail" can not be listened to, because. they "buried" the mail many times and so far without success.
Skype for Business (SfB) (formerly Lync) is a corporate messenger with advanced features. It has tight integration with the Exchange server, but is significantly inferior to the latter in terms of popularity. Skype for Business is usually used only in large companies, because. It is not very interesting for small and medium-sized companies.
The current version is Skype for Business 2019, which has minimal differences compared to the previous version of Skype for Business 2016, so SfB 2019 can be conditionally considered Service Pack 1 for SfB 2016, and not a new full version.
In the Office 365 cloud, this product was represented by the Skype for Business Online service, which after some time was completely replaced by Microsoft Teams, i.e. Skype for Business is not currently available in the Office 365 cloud. For this reason, it is hardly worth expecting the next local version of Skype for Business 2022, since the priority for Microsoft is the development and development of the Teams messenger, which has become the vendor's response to the emergence of the successful Slack messenger.
If you currently administer a local Skype for Business and you like the concept of a corporate messenger, then I advise you to look at Teams as part of Office 365, otherwise it is better to choose another product to upgrade your knowledge, because. oblivion awaits local Skype for Business. Unlike Exchange, which has become the de facto standard in the mail server niche, Skype for Business today has alternatives. Team and Slack for large and medium companies. Telegram, Viber, Whatsapp - for small companies.
SharePoint - an internal corporate portal where companies can post their useful Web services (vacation schedule, list of employees with photos and phone numbers, birthday reminders, corporate news, etc.). Users can store, edit, and share the files they host in their SharePoint libraries.
SharePoint is like Bitrix24, only bigger, more functional, more expensive and much more difficult to set up and maintain. The killer feature is the ability to simultaneously edit a single document by a large number of employees, which is very convenient when 100 people are trying to fill a vacation schedule, and integration with Office Online Server and local MS Office.
Sharepoint is a large, complex, and expensive product, so it's usually only used by large companies. Smaller companies use Bitrix24 or its analogues, or simply store files on file servers, and distribute useful Web services to different internal sites.
SharePoint farms (clusters) are usually managed by developers with administrator functions, and not by "pure" system administrators, because in order for SharePoint to take off and be useful to the company, it needs to be finished with a lot of code.
Office 365 includes the SharePoint Online service, which is a simplified version of on-premises SharePoint, i.e. has a minimum number of possibilities for customization and βdrank it up for itselfβ, but it removes a lot of headaches from the developer administrator regarding its operation. My verdict is that the complexity and high cost of maintaining on-premise versions of SharePoint will do their job and companies will happily begin to gradually crawl into SharePoint Online, or abandon Sharepoint altogether in favor of some simpler solution. I personally do not see a bright and carefree life for SharePoint in local installations.
System Center is a whole family of products for deploying, configuring, managing and monitoring large Windows infrastructures. Judging include: System Center Configuration Manager (SCCM), System Center Virtual Machine Manager (SCVMM), System Center Operations Manager (SCOM), System Center Data Protection Manager (SCDPM), System Center Service Manager (SCSM), System Center Orchestrator (SCORCH ).
The full range of System Center products is usually in demand only in large companies, while medium-sized companies usually use only one or two products.
Since System Center products are quite difficult to learn and are usually used only in large infrastructures, it is customary to allocate individual people to work with them, for example, a monitoring systems administrator (SCOM), a workstation maintenance administrator (SCCM), a virtualization systems administrator (Hyper -V + SCVMM), Infrastructure Automation Administrator (SCORCH + SCSM).
Microsoft is rapidly developing its cloud services, so System Center functionality is gradually moving to the clouds. All this will have a strong impact on on-premise System Center products in the very near future.
Functional System Center Orchestrator (SCORCH) will replace the Azure Automation service in the future ().
Functional System Center Operations Manager (SCOM) will replace the Azure Monitor service in the future ().
Functional System Center Data Protection Manager (SCDPM) will replace the Azure Backup service in the future ().
Functional System Center Service Manager (SCSM) will cease to be required or will be replaced by any other ticket system, for example, Jira.
System Center Virtual Machine Manager (SCVMM) remains with companies that run Hyper-V virtualization on-premises for now. Small Hyper-V installations (10-15 servers) can be quite successfully administered without SCVMM using only standard tools - Failover Cluster Manager, Hyper-V Manager, Windows Admin Center.
System Center Configuration Manager (SCCM) - used for mass deployment of operating systems, installation of corporate applications from a single directory, installation of Windows updates on servers and end workstations, application inventory and license counting. It seems that this is the only product from the entire System Center line that will remain with us in the on-premise infrastructure, because. it is not possible to fully replace it with something cloudy at the moment.
If you currently support a local installation of System Center Configuration Manager (SCCM), then you can continue to do so. the product will be with us for at least the next 3-5 years. Additionally, I would recommend starting to explore the possibilities of Office 365, because. this will go well with the position of Enterprise Desktop Administrator.
The administrator role for most other System Center products will be retired as Azure services significantly simplify their work, hiding all the complexity from prying eyes. Let's take an automation administrator (SCORCH + SCSM) as an example. SCORCH will be replaced by Azure Automation. Knowledge of the automation process, PowerShell, SQL will remain and will be useful for Azure Automation, but knowledge about building SCORCH clusters, ensuring their high availability, sizing resources, updating, migrating to new versions, backup and monitoring will lose their relevance, because. All this work will be taken over by the Azure cloud. The automation administrator will only focus on the automation process itself, as all the work to maintain the efficiency of the automation infrastructure will go away from him.
Windows server and its roles
Active Directory (AD) - a place where user and computer accounts are stored. If the company has more than 20 computers, then most likely there is already some kind of Active Directory domain there. Knowledge of Active Directory, the ability to distinguish a domain from a forest, the skill of working with group policies are mandatory for any Windows administrator. This knowledge will be relevant for another 20 years. Additionally, I would recommend that you familiarize yourself with Azure AD (AAD), look at options for synchronizing users between on-premise and cloud infrastructure.
DNS, DHCP - network services, the understanding of which is useful in all areas of IT, from administration to programming, so you need to know them. Understanding the operation of networks, routing protocols, OSI and TCPIP models will be a definite plus for any IT specialist.
Hyper-V - the name for the entire stack of virtualization technologies from Microsoft and its hypervisor in particular. It is developing quite rapidly, although in my opinion, most of the new features (Shielded VM, Encrypted Subnets, Storage Spaces Direct) are focused primarily on local (Cloud Service Providers) and global (Azure) cloud providers, and not on the corporate segment (Enterprise). This is generally understandable, since Microsoft first implements and tests new functionality in its Azure cloud, and only then transfers it to Windows Server and Hyper-V.
Hyper-V still suffers from the lack of a single free console that would provide all the necessary features. Now we have Failover Cluster Manager, Hyper-V Manager, Windows Admin Center. Such a console was supposed to be SCVMM, but it is paid and somewhat difficult to master.
If you currently support a local installation of Hyper-V without SCVMM, you can continue to do so. In parallel, I would recommend starting to study Azure IaaS and the mechanisms for migrating virtual machines between the cloud and on-premises infrastructure.
Among my environment (banks, telecoms, insurance companies, large industrial holdings), all productive virtualization is usually managed by VMware vSphere, and not Hyper-V with SCVMM, so I can recommend the Hyper-V administrator to also look towards VMware and its products .
Cloud services
Office 365 is a cloud service that provides a subscription package of Microsoft Office applications (on-premises and its Web versions), and also includes the main server products - Exchange, Teams, OneDrive and Sharepoint.
At the moment, Office 365 is a self-sufficient service that almost completely covers the needs for office communications. Due to the ease of setup, it is great for both small companies and medium and large businesses.
The presence in the cloud of already deployed services Exchange, Teams, OneDrive and Sharepoint significantly reduces the burden on the system administrator, because. all procedures for installation, resource sizing, updating and migration to new versions now lie entirely with Microsoft. If earlier it would take 4-6 separately dedicated administrators to maintain Exchange, Teams, OneDrive and Sharepoint in the local infrastructure, now only 365 average administrator is enough in Office 1. If something does not work or does not work correctly, then you can create a ticket to Microsoft technical support directly from the Office 365 interface, which is very convenient.
If you are now a system administrator maintaining on-premises versions of Exchange, Skype for Business or Sharepoint products, then I would recommend looking at their cloud versions as part of Office 365 to understand how they suit you and what functionality they provide compared to on-premises versions.
Azure is a global cloud platform from Microsoft that includes an ever-expanding set of cloud services to help organizations meet their business challenges. At the moment, Azure includes more than 300 services grouped into different categories (computing, networks, storage, databases, analytics, Internet of things, security, devOps, containers, etc.).
Having appeared for the first time in 2009, Microsoft Azure now occupies one of the leading positions in the global cloud services market, successfully competing with Amazon AWS there.
According to the latest financial report () quarterly (Q4 2019) profit of Microsoft grew by 49% due to the success of Office 365 and the cloud business. Revenue from Azure grew by 64%.
Azure, along with Office 365, are the main areas where Microsoft directs its financial and organizational resources.
The abundance of services on the Azure platform can confuse even an experienced IT specialist, so below will be a description of a typical server screw infrastructure, where in brackets I will indicate their approximate analogues in the Azure cloud. I hope this will serve as a starting point for learning about Azure, because, as you know, you need to start small, gradually moving deeper.
A typical server-side Windows infrastructure looks something like this:
- Active Directory (AD) with group policies and DNS. (Azure Active Directory (AAD), Azure DNS).
- DHCP
- Exchange mail server. (Exchange online as part of Office 365).
- RDS farm with multiple terminal servers. (Azure virtual machine + Azure Virtual Network + Azure Storage).
- A file server where employees store their files. (Azure File Storage, Azure virtual machine + Azure Virtual Network + Azure Storage)
- Servers with applications and databases (1C, internal website portal, CRM, etc.). (Azure SQL Database, Azure Web Sites, Microsoft Dynamics 365, Azure virtual machine + Azure Virtual Network + Azure Storage)
The main administrative tasks are:
- Creation of backup copies. (Azure Backup).
- Collection and analysis of logs. (Azure Log Analytics).
- Automation of routine tasks. (Azure Automation).
- Monitoring the status of services and receiving notifications of failures (Azure Monitor).
I would advise Windows administrators maintaining local infrastructures to first look for analogues of their favorite services in the Azure cloud in order to work with them a little, determine their usefulness for the company, and perhaps organize hybrid options, choosing the best of both worlds.
Training
The focus of Microsoft in the development of its products is gradually shifting to cloud solutions, so you need to start learning them now. Where can I get more knowledge about Azure in Russian? Unfortunately, there are not many such resources.
Microsoft offers to use its Microsoft Learn Portal - . The text material has been translated into Russian, while the video is in English, albeit with Russian subtitles.
As a good and high-quality material for studying Azure, I would recommend the Exam AZ-900 Azure Fundamentals course, which Igor Shastitko reads on his YouTube channel (). Now there are 13 videos, but if there is enough active support from the community (like, subscription), the materials will appear faster and the continuation will not be long in coming.
Additionally, on the iwalker2000 channel, I advise you to watch the playlist "IT Career: How to Become an IT Specialist", which will help beginners determine the path of their professional development and build a career in the right way. ()
Unfortunately, there are not as many materials on Azure in Russian as we would like, so if you know any other useful resources on this topic, please share them in the comments. Many IT professionals will thank you for this.
Conclusions
What conclusions can be drawn from all of the above?
- Life in the Microsoft infrastructure is still there, and it's not going anywhere. Microsoft has a fairly wide portfolio of software solutions, many of which are leaders in their niches, so a system administrator and engineer always has something to learn, implement, operate and develop.
- The Microsoft infrastructure is actively changing now, and this is happening with an emphasis on the development of cloud services - Azure and Office 365. New Microsoft products and applications will initially be created to work in the cloud, tied to a subscription model with monthly payments. Only some of these products will later find their embodiment in on-premise solutions.
- Some of the expensive and difficult to maintain products will soon leave us, moving in whole or in part to the Azure cloud or Office 365. Individual administrators who constantly maintain only 1 product (for example, SCOM, SCSM, etc.) will soon be abolished.
- If you are an experienced system administrator/engineer working in the Microsoft ecosystem, then it is not necessary to drop everything and run to DevOps, which is now being talked about on every corner. You can continue to develop further in your direction by adding competencies in Azure cloud services and Office 365.
- To remain in demand as a specialist in the labor market, you will have to study, study and study again. The concept of "lifelong education" for IT is more relevant than ever, especially now in times of active development of cloud technologies.
- DevOps is now at the peak of its popularity (hype). It is a fact. Initially, DevOps was perceived as a methodology that allows you to combine software development and operations, when programmers and engineers begin to work together for the benefit of a single common goal - to make software better. The main emphasis was placed on changing the culture of communication between teams, developing mechanisms for mutual assistance and collective responsibility for the final result. However, as a result, this led to the emergence of a new position - DevOps engineer, who was delegated the tasks of release engineer (CICD), automation administrator, cloud administrator and operations engineer. This is already a fait accompli. The number of vacancies for DevOps and the requirements in them only confirm this.
DevOps can now be seen as an additional path for the development of a system administrator/engineer. DevOps is a great way for the average administrator to change their current industry to the software development industry. Those who like automation and writing code will eventually become developers, and those who like infrastructure things (networks, servers, OS, clouds, etc.) will become DevOps engineers.
- If you are a beginner, or just entering IT, then DevOps is now a great way to upgrade in a short period of time and get a job in a normal company with decent pay and a good office, so learn Linux, Ansible, Docker, Kubernetes, Python and CICD .
Recently, the demand for the Linux platform and solutions related to software development has increased, but this is not happening at the expense of the Microsoft ecosystem, but simply a new niche has appeared where Docker and Kubernetes are actively used, monolithic applications are being cut into micro-services, and business needs to increase the speed of software releases to reduce the time to market for new features.
Source: habr.com