#KingstonCognate introduces Rob May
Rob May an award-winning speaker, delivers keynotes internationally and runs CEO and Director workshops for numerous business leadership organisations in Europe. He speaks as a current and very relevant expert being founder and Managing Director of ramsac, with a team of 75 consultants working with him (and an alliance partnership with PwC) to deliver IT and Cyber security services/support (and have done so for 29 years). Rob is the UK Ambassador for Cyber Security for the Institute of Directors and he is ranked No. 5 in the Global rankings for Cyber Security Thought Leaders/Influencers. He is on the Advisory Board of The Cyber Resilience Centre for the South East, working with industry, academia, and law enforcement.
As MD of ramsac, a Managed Service Provider (MSP) for nearly thirty years, I have seen an awful lot of challenges, we have had recessions, technology shifts, and pandemics, along with a backdrop of economic and political pressures.
Lockdowns in their various formats and durations have undoubtedly caused hardship and damage to many businesses but there are also businesses that are thriving in this current economy. I deal with a lot of businesses who are currently doing extremely well. The onset of the pandemic certainly changed working life for most people but on the other hand, some have experienced positives.
Technology support companies typically charge on a cost per user per month and many will have seen decreasing annuities because of supporting clients with furloughed staff. There have also been changing business practices and business models where demand has most certainly offset the decrease in other areas.
Today’s business culture
Today’s model of work has become remote and it has driven the need for upscaling cloud and data center provision. Many businesses are investing in technology, reconfiguring office space to reflect these new needs and relocating server infrastructure. These needs and demands have spurred increase in server upgrades (both on premise and in the cloud/data center). This is a sensible development, but I see many organisations getting it wrong.
Why do organisations upgrade?
A typical upgrade is going to comprise of both memory and storage upgrades.
A memory upgrade is the easiest way to improve and optimise server performance. It is usually required when increasing virtual workloads, adding more users, implementing a new database program, or perhaps when adding commerce functionality to a website. Upgrading memory in your servers can support larger workloads, provide faster responses for those using the cloud and improve speed for applications living in memory.
Increased server memory allows you to have more virtual machines and can increase allocation to maximise your existing hardware. Virtualisation lowers total cost of ownership and more memory for Virtual Desktop Infrastructure (VDI), which means a faster response when apps sit in memory. (VDI is desktop virtualisation where the desktop environments are hosted on a central server and delivered to end clients over a network).
IT engineer installing storage drive into server rack
Storage is an important component
Storage is important for server virtualisation in terms of capacity but most significantly, in terms of speed. Random input/output and high data access demands can be severely limited by traditional mechanical hard drives in a virtualised host server. Servicing applications, users and retrieving data have become limiting factors in overall system performance. Installing the right server memory along with Solid-State Drives (SSDs) which are non-mechanical, are the ideal accelerators for improving performance and capacity which allow for virtualisation and eventually being able to do more with less.
As server storage is increasingly offsite in data centers and is becoming more cloud-based, greater demands are placed on both the storage drives and the memory infrastructure. Solid-State Drives (SSDs) have fast-retrieving capabilities and enable cloud services to operate at optimum efficiency for each type of data being processed. A well-designed storage solution will deploy a tiered drive model (differing speed and price drives for different storage and access purposes) which ensures an organisation has the right data on the right tier to maximize performance while remaining economical.
Reliability and server performance are key
It starts at the component level and many people do not properly consider memory (which is a big part of the server price). Server DIMM memory modules are either going to be unbuffered uDIMMs, registered RDIMMs or load-reduced LRDIMMs. All three types are compatible physically, but you must specify the correct DIMMs as they CANNOT be mixed.
It’s also vital to match the correct memory to the exact processor you have in each server. The number of memory channels and memory modules can have a significant impact on performance. People need to keep this in mind when considering a staged upgrade. Do not buy 12 months of memory today for an upgrade that goes beyond your requirements or even limits your performance. Plan and make sure you get expert advice on the configuration for each individual server.
Memory standards and specifications are stipulated by an organisation called JEDEC who are globally responsible for the development of open standards for the microelectronics industry. Paying attention to their standards and specifications can help with informed decision making. For example, current considerations on motherboards really need to look at the new standard for DDR5 which dramatically increase memory density and speed. DDR5 DIMMs will have the same number of pins (288) as DDR4 DIMMs, but the pin layout is different and can only be used in a dedicated DDR5 slot.
Making the right choice
man points with his hand at an illustrated square in front of him with a green check mark in it
Upgrades are a shrewd decision in our ever-changing, ever-evolving business landscape. Planning, specifying and buying for the right upgrades, however, is an art! Get it wrong and you will fail to get optimum performance and waste time and money. Do not rely solely on the advice of your server manufacturer either, as in my experience, they will rarely take the time to work through your plans and/or look correctly at all the configuration/performance issues but rather try to sell you more memory. Personally, I would use the “Ask an Expert” team at Kingston who will be able to advise you on nearly all server platforms and their advice is rarely ever wrong!
Do not simply replace servers. Get good advice and look at an intelligent upgrade roadmap that covers your current needs and your projected growth.