Optimal office network advice

From home grid
Jump to: navigation, search

An optimal office network requires serious research & advice. This is VERY generic advice.

AMD AM4 vs intel[edit]

As of 2018-11, the AM4 platform had consolidated low-end office desktops for non-gaming uses, on the somewhat cheaper & more heat-robust AMD platform. A gigabit guarantee to local transit exchange could be assumed in most urban & small town settings where the business actually justifies it, so most LAN host functions had shifted to a cloud service, except where contracts or performance required otherwise. For instance, a law firm or medical facility have specialized legal privacy and medical privacy concerns. Removing the cost of such hosts makes it more feasible to invest in better desktop, laptop, tablet, phone, watch, projector, headsets, but at a longer term cost & breach risk to access the organization's data as it is stored remotely.

Very few users actually require intel's single thread performance. The price premium (huge) tends not to be worth it, as there is no reliability advantage especially not where heat failures are a problem (AMD is generally better at this). AMD's single socket type (AM4) makes it preferable to intel's maze of always-shifting sockets that disallow processor upgrades & make all intel-based networks less modular.

introduction: non-optimal office networks[edit]

Imagine a badly designed ad hoc network that slowly evolved over time. See optimal home network advice for bad example.

Maintaining such networks has unpredictable costs, but more important, it can waste weeks of its owners' time without notice. The lack of interchangeability requires a great deal of expertise and research to resolve. Even where expertise is available cheaply for typical problems like coolers or memory upgrades or disk swaps, it simply amplifies the heterogeneous nature of the network .

That said, properly run, with data stored in multiple places and boot images restored to the relatively fragile boot drives using bootp etc., such a network can probably gronk along.

"optimal" assumptions[edit]

The total cost of ownership for an optimal office should stay under $3000/year/worker for typical document-oriented office work but up to $8-10,000/year for heavily supported professional work like design or engineering drawings. It should not be much lower, as that entails risk (poor backups, etc.).

Initial outlay should never be less than half of that and no device with clearly <6 year anticipated lifespan should usually be bought. That includes mobile devices. Network needs will vary with tasks.

An employee doing work for hire should be required to understand at least how to report a problem [1] and engaged in constant improvement to terminology that generally becomes more practical as stable systems become more practical. Less sophisticated users should rely on more sophisticated peers, and equipment should be upgraded and standardized no less than every 4 years (with devices at the end of their lifespan allocated to unique tasks, or sold off as they will have >2 years lifespan left).

not for gamers or Apple users[edit]

This advice is not for those whose primary needs are gaming or involvement in the game industry or its development, nor for those using Apple equipment, which uses its own protocols, cloud service, etc. An

An iPhone or iPad should be supported as an exception to a network generally built with Linux back end and Android & Windows 10 front ends in 2018, as MacBooks & iMacs are very much rarer & hard to support.

Assume expensive power[edit]

Businesses as a rule pay more for electricity: A maintenance price per year should assume at least $0.20/kwH electricity as a North American average with $1.00/kwH on-peak for the period 2018-2025 or so. Properly future-proofed devices should remain performance-per-watt competitive to run for 8 years minimum, assuming expensive power may allow 10 or 12 years. A separate emissions calculation may have to be made, especially where there is a carbon tax that affects electricity pricing making on-peak higher.

Since <17W processor+graphics combinations are available (using laptop technology) on the terminal and other power-conserving infrastructure is freely available, it's wise to minimize wall warts from DC devices. IEEE 1901 should eliminate most AC waste from the smarter devices that use it. Innovating in these areas is justifiable for spreadsheet-wise corporations, and increasingly working in a power-cost-sensitive, heat-sensitive, green-ness-sensitive, business environment, where there may be healthy telecom infrastructure concerns.

Jurisdictions in decline, like Ontario or New Brunswick, have directly subsidized expensive dirty (nuclear, dam) consumption often at a cost >$100/resident/year, much more when subsidies to generation (including "green" generation) are included. Unsurprisingly such jurisdictions face a carbon tax & emissions boycott problem in 2019-25. Dirty power cannot become or remain cheap without permanently destroying most prospects in/for the new/green/efficient economy. Reputations are very hard to restore.

model transactions/s, transactions/joule, transactions/s/$[edit]

If large scale processing is happening, this will require some specific modelling for that industry.

initial cash outlay[edit]

miniITX AMD4 platform, supported through 2020 with new processors, some of which are very cost-effective (Ryzen 1700)

Monitors should be 40" or less 2160p or "4K", or UHD (1440p). 1080p is acceptable for video & for video projectors (only), 720p for marginal uses. These will actually be the largest expense & should be expected to last 6-12 years. Larger 4K TVs are desirable for high end presentations only, so should be deployed in boardrooms. Anything above 43" is unsuitable for use as a desktop 4K monitor.

A 2018 office network machine should put >288GB of SSD+RAM (of which at least 5% must be DDR4-1600 or better) within PCI-E x4 bus speed reach of a teraflop processor, with options to stripe that & double the speed again (using two PCI-E x4 slots perhaps sacrificing the ability for a 2nd video card).

This minimum should be sufficient for the combination of all boot/swap/core applications. The inherent fragility of striping requires a "wipe and re-image" mentality similar to a public library or government public access terminal - all boots installed on the SSD must be thought of as in RAM.

LAN host[edit]

Three or more terminals in heavy use with high value work require central support. A LAN host must be connected via at last one preferably two GbE or one 5GbE or 10GbE copper wires: ganging these into a 10 to 20gbps connection (minus overhead) is adequate if terminals have 16GB+ RAM. Note: wireless connectivity is NOT a substitute for wired, there are many failure modes of wireless networks that apply.

This host should cost about $1200 with one socket in use, $1500 with two (including cost of dual socket mainboard) and configured with 32GB of DDR4 RAM (maximized on build). By no means should hardware in any LAN host box (on which many users rely including usually employers of teleworkers) change lightly. It would ideally never change at all. Like a router, the value of such a box is absolute stability over an 8 to 12 year lifecycle. A list of reasonable upgrades, preferably all performed at once halfway through its lifecycle, is listed below.

terminals with 288GB-528GB[edit]

Terminals should avoid spinning metal storage. As of 2018 PCI-E x4 SSD with good controller was < $0.5/GB in 100GB-ish configurations. Onboard SATA6 is suitable only for replicated data storage. It must be backed up at least daily (for work, continuous) to a host NAS.

how to upgrade terminals[edit]

Any need for more or faster local storage requires a hybrid ssd on the terminal or SATA6 RAId. This should be the only use made of the SATA6 interface with extreme caution taken re use of raid (having boxes with identical mainboards & configurations onsite is essential as a dead RAid cannot be accessed any other way).

USB3 is strictly for transitory mobile removable drives/devices. If the terminal supports power over Ethernet this will allow support of VoIP & other extremely useful desktop & mobile support devices like a wireless AP. This should obviate any perceived "need" for Wi-Fi transceivers in the box, or (worse) USB networking.

LAN host[edit]

you will find it necessary sooner rather than later to segment public vs. private networks, dual-host most boxes and add a LAN host on the internal segment, removing file service & backup duties from the router (which will continue to face the slow open Internet segment or side).

A LAN host can do both duties with a trusted well-managed firewall, relegating the router to wireless connectivity only and freeing it to be located in the optimal wireless networking spot (rather than at the nexus of wires).

Ubuntu Linux is as of 2018 still likely the best OS for such a LAN host, in either role. If Linux is the primary OS for the terminals then it should also be used, for simplicity, on a self-maintained LAN host. Similarly if Windows Server or OSX experts are at hand, those can be used as well. As end users do not typically interact with this OS, it may be a matter of taste for the LAN administrator to choose a particular OS for such host(s). BSD is an option but OpenZFS makes it less necessary.

main host[edit]

Rackmount Xeon is still the preferred platform for ultra-reliable boxes. for some reason

storage: keep the spinning metal centralized[edit]

Spinning metal should be concentrated in the LAN host & main, and should not be used otherwise except for bulk media storage & backup.

what to upgrade, preferably all at once[edit]

Like a router, the value of a LAN host is absolute stability over an 8 to 12 year lifecycle. Only the following safe upgrade paths should usually be considered:

  • Add a second processor (when cheap) to a dual socket board - be careful, there are few insert cycles on modern chips...
  • Add second PCI-E ssd (when cheap), or two, or swap to liberate one to equip another terminal with the identical model as was used in the earlier terminals i.e. as spare)
  • (at end of lifecycle, or on major reconfiguration or failure)
    • add graphics PCI-E x16 & repurpose as workstation (if performance per watt allows) or display support system for less-used room (boardroom) if flops/watt poorer
    • remove excess storage & repurpose as router only
    • remove excess networking & reporpose as NAS only