May 23, 2012 at 22:54Network Hardware Update
The network works, 1C opens, users are satisfied. The picture is encountered all the time. It seems that the administrator’s life was a success. That is exactly what half, if not two thirds of novice and fairly advanced system administrators think. Many of us do not even think about what is in the network cabinets in the office, at best controlling what is in the server room. This approach is especially dangerous for your work (get fired) and for your organization (gets up). If everything works, it's time to update the equipment.
It is amazing, but often the opening of a network cabinet occurs at the very moment when this or that technology urgently needs to be introduced, but its implementation is slowed down or becomes impossible precisely because of outdated network equipment. Worse, but not less often, network cabinets open in search of a burned-out media converter \ switch \ access point when the work of a whole department, and sometimes the whole company, rises.
How to understand that the network equipment in the network requires updating and when to start it? How to approach the issue?
The answer is simple - you need to update the equipment constantly, cyclically, with a margin. However, it all depends on your company, its size, scope of its activities, prospects for its growth and growth of the local network, which is not always interconnected, used and planned to use network technologies, budget, approach to allocating funds for IT and hundreds of other factors . I offer you several criteria that can help in understanding the need to upgrade network equipment and the rationale for updating the owner.
Security can be internal and external. External security usually starts with a router, internal with a switch and access points. If the issue of protecting access points is often discussed, and the baby knows about protection from the Internet, then few people think about protecting LAN ports. Answer the following questions:
Test questions:
A list of questions:
The zoo is a great place to spend a weekend, especially now, in the spring. If you love the zoo, then you will not be confused by the presence of D-Link, Planet and Cisco in one closet. In the general case, such confusion should be avoided. The fact is that it is difficult to own different firmware of different manufacturers equally well and quickly. There is always a chance at the zoo to stumble upon various implementations of one standard by vendors. I strongly recommend that you devote a lot of time to selecting the right equipment from one manufacturer and checking compatibility.
A budget is one of the criteria for the need and speed of network updates. You are a system administrator. You must not solve problems, but warn them. You must inform your management of the current state of the infrastructure and the problem areas that exist. This approach will protect you from sudden claims. If you applied for equipment and justified its necessity, then you can safely refuse to introduce this or that technology in a handicraft way before acquiring the necessary equipment.
High-quality network equipment costs serious money. Cost planning will raise your level in the eyes of management and make your work easier. If you plan to purchase 1-2 switches per month, you only need to configure exactly 1-2 devices for this month. This is a very smooth and unhurried job, which allows you to choose the optimal time to replace the device, before that leaving you a lot of time to test and configure it.
The situation looks completely different when you, for one reason or another, need to update an entire network segment, for example from 5-6 end devices. Such a pace involves connecting more than 1 device per week, which does not always allow you to properly test and configure and almost never allows you to choose the optimal time for the physical replacement of the device. Those. quick replacement of devices either interrupts the work of some employees, or makes the IT department work after hours (days off / after the end of the working day).
From the point of view of allocation of funds, everything looks as follows: In the case of a planned equipment upgrade, this is a planned expenditure of XX rubles for the UE months, which is budgeted along with the purchase of an office. In the case of updating equipment for the implementation of technology - this amount becomes the amount of technology implementation. In turn, this can make the company abandon promising technology (ZZ), since freezing UU * XX + ZZ rubles can immediately be fatal for the main business.
A simple example. IP video surveillance wins compared to analog in many respects due to savings on laying kilometers of wires. This works if the IP camera connection infrastructure is ready. If there is no infrastructure, the cost of IP video surveillance becomes insane even for large enterprises. This ultimately leads to a double, triple and even quadruple waste of money.
First, they spend money on installing analog video surveillance. Then to install normal network equipment in connection with the failure of the old one or the introduction of IP-telephony. Then, to cover the damage from the emergency, since the culprit’s face is not visible on the analog recording, or nothing at all is visible, because there was a thunderstorm and everything is in the noise, then on the installation of digital video surveillance.
So, we discussed the main reasons for updating network equipment. Naturally, there are many other reasons, and each system administrator who has gone through this will call his own. I am sure that a lot of such recommendations will appear in the comments.
The described task for many seems impossible. Knock money out of the blue? Difficult, but necessary ! First of all, to you and your company.
It all starts with a plan. I offer you a standard plan that will allow you to understand where to start.
That's all.
It is amazing, but often the opening of a network cabinet occurs at the very moment when this or that technology urgently needs to be introduced, but its implementation is slowed down or becomes impossible precisely because of outdated network equipment. Worse, but not less often, network cabinets open in search of a burned-out media converter \ switch \ access point when the work of a whole department, and sometimes the whole company, rises.
How to understand that the network equipment in the network requires updating and when to start it? How to approach the issue?
The answer is simple - you need to update the equipment constantly, cyclically, with a margin. However, it all depends on your company, its size, scope of its activities, prospects for its growth and growth of the local network, which is not always interconnected, used and planned to use network technologies, budget, approach to allocating funds for IT and hundreds of other factors . I offer you several criteria that can help in understanding the need to upgrade network equipment and the rationale for updating the owner.
The first criterion: security
Security can be internal and external. External security usually starts with a router, internal with a switch and access points. If the issue of protecting access points is often discussed, and the baby knows about protection from the Internet, then few people think about protecting LAN ports. Answer the following questions:
- What can an outsider with a laptop, connected to the LAN port of an accountant's computer and knowing some initial data about the network, be able to do?
- How likely is such a connection?
- Who will be affected by his attack?
Second Criterion: Reliability
Test questions:
- How old is the existing equipment? You can look at purchase invoices or by production date if documents are not saved.
- How many years is it designed? Any equipment has a mean time between failures.
- What are the losses of the company from its sudden failure? Here you need to assess the number of connected people (including transfer) and the cost of an hour of their work.
- What is the replacement period and procedure? It’s one thing to take a reserve from a shelf. Another thing is to run through all the approvals with an account for a new piece of iron, wait 3 days for the vendor to receive the money, and another week for the piece of iron to be delivered.
- In what conditions does the equipment work? A server cabinet is one thing, a dusty ceiling is another.
- How much equipment is on the network? Let's say you have a floor with 20 users, the entire network infrastructure is reduced to a single cabinet, in which there are \ lies three switches with 8 ports. Will such a scheme be reliable?
Third criterion: load and potential
A list of questions:
- How loaded is this equipment? It’s one thing if you have a gigabit network, there is no IP video surveillance, there is no IP telephony and all the switches are connected to one switch core. I can envy you, but in 70% of small and medium-sized enterprises the situation is far from described.
- Are there any levers to control the download? You have a fully managed gigabit network and you just need to configure QoS - you are almost on the beach and can think about summer. Do you have a 100 megabyte network that has not heard of QoS? In the near future, the introduction of IP-telephony and IP-video surveillance? You will soon be hot, you are on your way to the hell of hell.
- Are there any levers for increasing throughput? It’s good if you have the opportunity to drag 100 meters of cable and connect a 100 megabit switch with its gigabit port to the core of the network. The situation is completely different if your 100 megabit switch with its gigabit ports can only be connected to another 100 megabit switch, without gigabit uplinks.
- Will an increase in throughput be required in the coming year, three years, five years? If you have not implemented IP-telephony and IP-video surveillance, then you can assume that you will need, of course, to talk about a 100 megabit network. Although in a gigabit network with certain configurations and the number of ports, such a need may arise. Even in the absence of plans for the introduction of IP-technologies, it is worth thinking about gigabit - the amount of transmitted data is growing steadily.
- Does your equipment support PoE? Many do not think about this issue, but answer yourself, is it easier to drag a power cable to an IP camera from an outlet or power it from a switch? Of course, you can put heels of splitters in the network cabinet, but imagine what the temperature in the cabinet and the level of clutter will be. In conclusion, I note that in modern offices PoE can power: Wi-FI points, IP cameras, IP phones, readers and sensors of access control systems, and even a wall clock.
Fourth criterion: equipment manufacturer
The zoo is a great place to spend a weekend, especially now, in the spring. If you love the zoo, then you will not be confused by the presence of D-Link, Planet and Cisco in one closet. In the general case, such confusion should be avoided. The fact is that it is difficult to own different firmware of different manufacturers equally well and quickly. There is always a chance at the zoo to stumble upon various implementations of one standard by vendors. I strongly recommend that you devote a lot of time to selecting the right equipment from one manufacturer and checking compatibility.
Fifth Criterion: Budget
A budget is one of the criteria for the need and speed of network updates. You are a system administrator. You must not solve problems, but warn them. You must inform your management of the current state of the infrastructure and the problem areas that exist. This approach will protect you from sudden claims. If you applied for equipment and justified its necessity, then you can safely refuse to introduce this or that technology in a handicraft way before acquiring the necessary equipment.
High-quality network equipment costs serious money. Cost planning will raise your level in the eyes of management and make your work easier. If you plan to purchase 1-2 switches per month, you only need to configure exactly 1-2 devices for this month. This is a very smooth and unhurried job, which allows you to choose the optimal time to replace the device, before that leaving you a lot of time to test and configure it.
The situation looks completely different when you, for one reason or another, need to update an entire network segment, for example from 5-6 end devices. Such a pace involves connecting more than 1 device per week, which does not always allow you to properly test and configure and almost never allows you to choose the optimal time for the physical replacement of the device. Those. quick replacement of devices either interrupts the work of some employees, or makes the IT department work after hours (days off / after the end of the working day).
From the point of view of allocation of funds, everything looks as follows: In the case of a planned equipment upgrade, this is a planned expenditure of XX rubles for the UE months, which is budgeted along with the purchase of an office. In the case of updating equipment for the implementation of technology - this amount becomes the amount of technology implementation. In turn, this can make the company abandon promising technology (ZZ), since freezing UU * XX + ZZ rubles can immediately be fatal for the main business.
A simple example. IP video surveillance wins compared to analog in many respects due to savings on laying kilometers of wires. This works if the IP camera connection infrastructure is ready. If there is no infrastructure, the cost of IP video surveillance becomes insane even for large enterprises. This ultimately leads to a double, triple and even quadruple waste of money.
First, they spend money on installing analog video surveillance. Then to install normal network equipment in connection with the failure of the old one or the introduction of IP-telephony. Then, to cover the damage from the emergency, since the culprit’s face is not visible on the analog recording, or nothing at all is visible, because there was a thunderstorm and everything is in the noise, then on the installation of digital video surveillance.
So, we discussed the main reasons for updating network equipment. Naturally, there are many other reasons, and each system administrator who has gone through this will call his own. I am sure that a lot of such recommendations will appear in the comments.
We plan our actions
The described task for many seems impossible. Knock money out of the blue? Difficult, but necessary ! First of all, to you and your company.
It all starts with a plan. I offer you a standard plan that will allow you to understand where to start.
- We make a network map. Which switch is connected to which channel.
- We inventory the switches. How many ports \ how many current users \ how many potential users. Type of switch and its capabilities.
- We learn the management plans for the development of the company, the use of IT technologies, the existence of this office.
- We select equipment with a margin. Here you need to select equipment with a margin of capabilities and ports. It is desirable that the newest possible, so that its release does not stop three months after the start of the project.
- We are preparing a plan for laying new wires to the core of the network (we also calculate the cost and optics too!).
- We are preparing a plan for the gradual replacement of equipment. The plan should look something like this:
- January - replacement of the central switch. Pros: improving the reliability of all networks and access to the server.
- February - replacing the server switch, installing a spare switch. Pros: improving the reliability of servers, eliminating the point of failure by using hot-swappable.
- March - installation of a new router. Pros: improving network security from external attacks.
- April - installation of a new switch on the 2nd floor. Pros: separation of the "unit name" in a separate vlan.
- Etc.
- We give away plans for examination to acquaintances, either to a forum or to an audit company to prevent childhood errors that are often not taken into account because of attachment to the conditions of a particular company, rather than standards.
- We get a conclusion from independent experts in one form or another. We adjust our plans.
- We prepare a report to the management and receive either approval or refusal. In the case of approval - we increase our professional level, in the event of failure - we take responsibility for failures of network equipment.
That's all.