The New Year holidays are a great opportunity to sit down, reflect on the year, and the year 2020 is also a chance to take stock of the past 5/10 years and write an article titled "What I would advise my past self 5-10 years ago, and what today looks like."
KDPV (メモリーズ)
Under the cut: boring details, scattered sclerosis, unproven claims, and more.
Additionally:
- The graveyard of IT professions.
- This groan of ours is called a song — HR policy, recruitment, and the situation with staffing.
- Average salaries in the industry.
- Do mid-level developers dream of electric tractors?
- In summary — what would I advise my past self 5-10 years ago, and what about today?
TLDR: You should go into development, those who talk about tractors are not on the road (those who left don’t talk), there truly are no cheap hires, you need to learn English and OOP.
Instead of a preface:
And previously there was a joke about "noble young ladies," that they are loaded with knowledge before exams, and during the exams, they spill it all out, leaving nothing in the barrel. Unfortunately, this is not an exception anymore, but a rule. The instability of knowledge is viewed as normal. It reaches the absurd. In the senior years of the Foreign Languages department, students seriously ask their teachers if they need to know the words they learn in the final years for the state exam, or if they should also remember the words learned in the earlier years.
A.A. Lyubishchev. On the state of secondary education, 1956.
1. The graveyard of IT professions
Of course, many remember 2019 for the outcome with Rambler, but that issue has been discussed enough already. Personally, I remember this year as the year of a sharp increase in demand for 'DevOps' and 'DevOps engineers'. Meanwhile, responsibilities in job postings are still written as before, by copying from 2-3 job ads — developer, DBA, Linux administrator, for testing, and sometimes the one responsible for information security. It’s simpler to call the resulting position DevTestProdSecEtcAllOther than to try to untangle the web of desires.
It's strange that the tasks have not yet included finding clients and purchasing equipment.
Behind this hype, the following went unnoticed:
1.1 Further growth of clouds, starting with the big five AWS — Azure — Google — IBM — Oracle.
1.2 Falling SSD prices, perhaps already below the price of 10k disks of the same volume.
1.3 Active promotion of All as Software — Software-defined storage (SDS), Software-Defined Networking (SDN), Software-defined data center.
1.1 Further growth of clouds
The inevitable result of the first point has been discussed multiple times — the overall demand for "localhost system administrators" is decreasing significantly. While 10 years ago it was necessary for a sysadmin to have at least some understanding of server hardware, this is no longer required; now, it's enough to pay the bills on time, and everything else will take care of itself, in the spirit of Omnisys in the cloud.
Yes, the configuration of 'it's good if the servers are at least Supermicro and all the software is obtained from publicly blocked sources' is still unmatched in terms of cost/quality, but only until the largest expenses become personnel salaries. If the salary of an additional employee exceeds the cost of cloud services, the services will move to the cloud, and tasks related to cloud access will be outsourced. Pure business, nothing personal. In the local segment, help with document formatting, cartridge replacements, and of course, writing complaints to vendor channels will remain.
Summary — The profession 'in general' is gradually moving in the same direction as engineers for Novell, Lotus, OS/2, AppleTalk, and so on.
1.2 Decline in SSD Prices
Whereas in the past, solving data access speed issues required some computation, the solution now is simple — 'Get an SSD.' The result is impressive articles on Habr, in the style of
thanks to the transition to storage with All-Flash architecture, the performance of a rather critical system has significantly improved. We did not conduct precise measurements, but it feels like the speed increase was 50-70%
Measuring increases based on feelings is fresh, original, and very engineering-oriented. It seems simple — open the statistics 'before' and 'after,' but no.
How does this affect the market? Simply.
- There's no need to calculate (even with a calculator), just buy SSD disks (or even an All Flash array) and press 'on.' After that, the storage system will take care of everything. Completely.
- There is no need to maintain a magical storage system, which means a dedicated employee or (a separate position in the outsourcing contract) is not required. Once set up, we only need to replace disks and Ethernet/FC SFP modules. The main thing is not to replace the FC8 SFP with a 10GE (real cases from practice). However, even if the client insists on having a certified engineer from %Vendor% on staff — after all, that certification can be purchased.
Summary — the demand for storage engineers is declining, the profession of 'storage engineer' is losing the ability to draw talent from those who know and require less training, 'staff' from relatively small organizations. Dedicated storage engineers remain with vendors, integrators, and large enterprises, and even there, they will soon need to be grown from scratch. Meanwhile, young professionals should immediately think about their prospects, like the dog Sharik from the joke — how many years it will take to move to mid-level, and maybe it’s better to take a course in C#?
1.3 Active Promotion of All as Software
Over the last 10-15 years, 'everything as software' has grown from 'just virtual machines' to solutions like
, and ,
from to .
SDS has evolved from the level of 'just giving space via iSCSI' to large distributed systems (albeit with their limitations, of course — something many users of Rosreestr have witnessed).
Despite the fact that in local installations of 100 workplaces and 10 virtual servers such a level of virtualization is not required (not to mention the cost of the solution), the result will be the same. — if we start to configure something massively through software (automate, yes), then the number of required employees decreases.
However, I am calm about network engineers — in all somewhat large infrastructures, a software infrastructure for managing networks has already been deployed, and it won’t be possible to have fewer than that until AI solutions for network management come to market based on
However, as services begin to migrate to the cloud, the concept of 'calculating how much a Software-defined data center will cost us compared to its own' becomes increasingly interesting, and allocating not just virtual machines but a whole SDDC to a certain department (development?) may become financially justified.
The conclusion is the same. The transition from hardware to Software-defined reduces the need for skilled personnel in the primary layer "in pieces." This is largely because anyone can press buttons in a Web GUI.
2. This lament of ours is called a song — staffing policy, hiring, and the situation with personnel.
— Dad, what are we going to eat today?
— Nothing, I'm working on an interesting project with a friendly team.
Before reading this part, it's important to understand the following:
* Even the simplest form-shaper and JSON writer is at the center of production. They create added value (which, of course, must still be sold or accepted by act). This is "profitable IT."
* Any kind of system administrator (except when the integrator rents them out at $10 an hour, with $50 for a site visit, and there's a friend — a business analyst. Data Center, Power BI, measuring any of your metrics) is a cost center, consuming money and incurring unavoidable expenses. Such IT does not generate any net profit for the end customer.
Thus, any developer has many more advantages — they are a unit that brings in profit. This is clearly seen when comparing two stories on YouTube — the ITGM report #14 (presented by Old Hrych) and "The Maximum Salary of a System Administrator" by Ilya Rudi.
As a consequence of all of the above (including the series of articles "Why Didn't They Call Me Back"), often heard laments remind me of a well-known joke:Look to the left… - mountains…, look to the right… - mountains…, and if you poke yourself in the leg — it hurts!«.
Questions from HR sometimes become more astonishing:
— We are looking for staff, why did you not respond to our spam mailing 'to everyone', where are your soft skills?
— Why don't you want to work for a gray salary below your current white one?
— How can you not be pleased that our director bought a new Bentley? Where's your team spirit?
— What does Labor Code of the Russian Federation mean? First of all, we are a startup; secondly, we know better what is written in the Labor Code. Why go to court, we were communicating just fine.
— Why are you asking if employees have had heart attacks at work?
--- We have order and established business processes, namely a strict work schedule, dress code, accounting of entry and exit times, deduction of hours, multiplication of KPIs (and bonuses) by zero for being late by 10 seconds, budget approval in just six months, and issuing a laptop in only one month — we are just waiting for you to bring order to everything else.
--- We are not Facebook, Amazon, Apple, Netflix, Google (FAANG) — but you should come to us. We don't really know why, maybe for the chaotic atmosphere of a startup! We have no money, but the startup will definitely succeed.
Conclusion: When someone claims that "there are no specialists", they are absolutely right, but the phrase needs to be read in its full version: "our HR cannot find candidates for the salary and conditions we offer, with the required knowledge. They do not have the desire or courage to say that our requirements do not match the market, and we do not check whether they are actually looking for someone or just sending messages and calling everyone on the lists of open vacancies and keywords. We are fine with this situation because we can always blame HR for the problems. HR is also fine with the situation as their KPIs are tied to the number of emails sent and resumes forwarded to us, rather than filling positions.»
It's a Win-Win situation.
Thus, the whines and cries of "oh, there are no specialists" from management and tales of "the interview again asked about the slots and how many bits a line took in Java 5.0 — then they asked whether I was confident in my qualifications ( ) will continue.
The issue is that
- The era of remnants of at least Soviet education has passed, and there are no more masses of cheap qualified personnel, and there won't be anymore.
- Candidates from the regions have been snatched up, and it is becoming increasingly difficult to engage the remaining ones; they are already working for rates of $15/hour and above.
- The demographic pit is no joke in statistics.
- Savings on IT constantly turn into a couple of days of downtime in the whole IT department or "our only key specialist has left, our system is in danger."
3. Average salaries in the industry
A brief study of Habr and hh showed that there is no significant volume of salary statistics. The published statistics from my circle — about 800-1000 vacancies in the field? (Examples: , ) Something certainly shows.
The salary database from hh is only available by subscription.
An attempt to analyze hh — "» by Ilya Rudya suffers from certain shortcomings — it only considers job vacancies with a 'salary' and only in the Windows direction.
Job vacancies in Telegram channels (and everywhere) suffer from salary ranges like '50-150' or '130-330' thousand rubles net, and increasingly, the actual salary plus bonus plus bonus criteria can only be known after 1-3 months of work.
The situation is especially complicated with the 'annual bonus' — it may not be given with the explanation 'the year was difficult, but you hold on.' This leads to significant savings on the payroll fund if the annual bonus amounts to 3-6 monthly salaries and only with it the salary is approximately equal to the 'market' rate, or rather 'the one promised in the neighboring organization.'
This situation has been stable for years and fully satisfies both employees and employers — I can't explain the absence of a Glassdoor analogue in Russia in any other way.
This same situation creates unreasonable expectations among newcomers in IT, where they expect salaries starting from *3 times the average in the city. For example, if the average salary reported in the press is 30,000 rubles, then anything less than 90,000 'net' is not expected. For Moscow, the 'average' ranges from 55,000 to 80,000 rubles, depending on which service calculated it and for what purposes.
In my assessment, reality is much worse; however, I stopped gathering statistics a year ago for the reasons mentioned above.
Reality for Moscow, estimated and not provable:
Any junior: 35-55 thousand rubles. With knowledge of English: 45-75 thousand.
Any middle: 50-150 k. Knowledge of English is assumed, at least at the level of let me speak from my heart, ?
further serious discrepancies in salary levels begin, depending on the direction, actual level, experience, and a hundred other reasons.
4. Do mid-levels dream of an electric tractor?
Citizens! Life dictates its own laws, its cruel laws. I will not speak to you about the purpose of our gathering — you know it well. The purpose is sacred. From everywhere we hear groans. From all corners of our vast country, cries for help arise. We must extend a helping hand, and we will do so. Some of you serve and eat bread with butter, while others practice their trade and eat sandwiches with caviar. Both groups sleep in their beds and cover themselves with warm blankets.
Discussing the topic 'Piglets — to the tractor' is undoubtedly an interesting undertaking, but it depends on several factors, namely:
— What exactly the opponent read. If it was about the Horde, where factories, steamships, and rockets are being built one after another — then of course, why go anywhere.
If, however, it happens in a competitive environment, the question arises as to why the opponent is still here after five years if things are so bad.
There is no publicly available statistical database for the numerical analysis of the situation regarding staff turnover.
The numbers of vacancies and salaries on hh show nothing, as vacancies are published three times, and the salary ranges are:
a) nonexistent
b) published as 'from 100,000 to 250,000 rubles'
c) after the interview, they might be LOWER than the initially stated range.
The data from the State Statistics Committee on the number of people who left and arrived shows only general statistics, like 'two million came from the southern countries of the former USSR, while 10,000 went to Europe and the USA, and it seems 9.5 came back'. Some reliability regarding the 'total number of those who left' is provided by the statistics of the receiving side, but it does not break down how many engineers in automated control systems left, how many developers, how many students, and how many children and relatives of respected people went 'to study in Europe and keep an eye on the vineyards'.
We can try to rely on survey data such as Currently, 53 percent of citizens aged 18 to 24 years would like to emigrate from Russia.. A great survey format; one could just as well ask 'would you like to receive a million', without clarification — a million dollars in small bills with no obligations or a million rubles at 40% per annum.
We can try to look for lottery data
There could have been a table here, but over the years, the platform hasn't managed to create a button for inserting tables, and I find it tedious to do it manually, especially since they can be glitchy.
Russia, total
2010 — 101,324
2011 — 140,444
2012 — 167,600
2013 — 218,862
2014 — 249,670
2015 — 265,086
2016 — 274,746
2017 — 332,069
2018 — 434,353
I don't know what conclusion to draw or how to read this data correctly. In any case, 430,000 applications for a green card correlate weakly with the 53% who are interested.
At the same time, as knowledgeable people irresponsibly state, juniors are not awaited 'there,' while 'seniors' are still well-fed here, perhaps even better. There is demand for middle-level positions 'there,' but the closer you are to senior, the greater the risk of a drop in living standards, going from marble steak back to instant noodles, since rent is expensive, healthcare is costly, and everything is alarmingly expensive.
Conclusion. There are dreams and conversations. However, there hasn’t been any publicly noticeable movement. Perhaps those who wanted and could have already left, or even returned.
(1) The competitive environment is like 'he who dines the girl, dances with her too'; more accurately, it's about what the administrator of the channel/group feels passionately about, and that will be the only valid point of view.
5. So, what would I advise myself 5 to 10 years ago, and what would it be today?
A table of pros and cons could have been inserted here, but after all these years, the platform still hasn't created a table insertion button. So it will be like this:
A developer is:
- Immediate profit for the business (A system administrator is considered a pure loss from the accounting perspective).
- A growing Russian and global market (The system administrator market is declining. After a dozen racks and with modern automation tools, the difference between 10-100 and 100-1000 is becoming less significant).
- The ability to work from home (System administrators sometimes have to go on-site).
- Global demand is increasing (for system administrators, it is decreasing; global clouds are often more advantageous than local ones due to their dynamism).
- Weak ties to existing and rarely updated infrastructure (What is absent from local infrastructure. Equipment lifespan is 3-5 years, it's not possible to switch instantly from Cisco to Arista or from HPE to any SDS, one cannot change physical infrastructure every day).
Conclusion. There are no longer any advantages to starting a career as a system administrator (ops) now, yet the requirements keep rising. Ops/SRE seems more appealing.
Note: meanwhile, in other sectors in Russia with an engineering focus, the salary situation is even more dismal. , therefore the prospects in SC (operation) are comparable to ASU TP, only there is Siemens instead of Cisco/HPE.
The advice remains the same as it was 5 and 10 years ago:
- Study programming theory and mathematics.
- In-demand languages change every 5 years, algorithms don't change as frequently.
- The second mandatory requirement is to learn English (optionally German or French as a second foreign language).
- Regularly read hiring advice sections for 'clients'. Topics like 'how to convince a candidate they don't deserve such a salary during the interview', 'how to convince an employee of the inevitability of bonuses and a soon promotion, or more precisely a change of title', 'stories about the benefits of working 12 hours and on Saturdays', and other career consulting courses and soft skills improvement from business veteranism.
As for forecasts, I think that 2020 will bring us a lot of new and amusing things under the banners of 'soft skills' and 'toxicity'.
Literature for extracurricular reading.
If someone has read this far, then:
THE STATE OF IT HIRING
In search of top IT talent
What IT executives are doing to fill critical positions and takeaways for job seekers.
Copyright 2018 Hewlett Packard Enterprise Development LP.
Candidate expectations can cause other wrinkles: 58 percent of respondents say they frequently or constantly run into candidates whose salary expectations are out of reach for their budgets.
It can be difficult to attract staffers with the coveted high-tech skill sets because good IT candidates are in demand across the industry—and they know it.
Special attention should be paid to page 10 — Sweetening the pot for potential IT hires.
Considering that global trends reach Russia in 3-5 years, by 2022, this document may be translated and made available to those interested.
Source: habr.com
