Those who are confident and absolutely convinced about Oracle Engineered Systems and its impacts (positives and negatives), please drop right away and save your time. This is not your cup of tea.
If you are a decision maker or an influencer who seek quick help to get the insights on purchasing Oracle Engineered Systems, this is for you.
This is not a marketing article for Oracle and I am not associated with Oracle or its affiliates in any manner.
If your company is a predominantly high consumer of Oracle software or hardware, it is impossible that a flamboyant oracle sales person hasn’t pitched to you with one or many flavors of its Engineered Systems stack. It’s an eventuality and a reality that you as a decision maker would face at some point in your tenure in the organization. In that case, how would you quickly gauge the proposed system, its merits or demerits, its ROI and TCO compared to traditional systems or public cloud alternatives?
This article or series of articles aim to help out those who want help in these. This is written as a series and this is the first of MANY in this.
Oracle Engineered Systems (OES)
In simple terms, Oracle Engineered System is a bundled platform which consists of network infrastructure, integrated cabling, server compute , storage system, and the underlying system and application software to orchestrate its infrastructure. For a reasonably tech-aware person, it is “ a finite number of servers attached to a finite number of storage servers, with disks of different performance levels attached to it , all of these are interconnected throough inbuilt infiniband and other kind of switches; and all of these things fitted in a 42 U Rack”.
Major engineered systems offered by Oracle are
- Exadata – Running Oracle Database
- Zero Data Loss Recovery Appliance – Running data backup, protection and data lifecycle management solution
- Private Cloud appliance – Running generic IaaS orchestrated by Oracle’s virtual machine software
- Database Appliance – Running oracle Database with inferior features compared to Exadata
- Bigdata Appliance – Special purpose platform for bigdata management
- Exalogic (Stopped production) – A special purpose system for Oracle Weblogic and Coherence
- Exalytics (Stopped production) – Special purpose system for analytics
- Super Cluster (Stopped production) – High performance compute for database and applications
Essentially, Oracle bundle together each piece of hardware required to run specific solution; like Database or any other and deliver the same on premise or in cloud. It’s “In-a-box” solution as claimed by Oracle.
“Why OES?” – ORACLE’s claim (offer)
Let me quote Oracle’s claim on its Engineered Systems.
- Reduce risk and lower costs
Oracle claim that, they have bundled together the best of breed hardware , software and firmware components together “in a box” . They claim that, they engineered , integrated and tested on every bit of it and hence it reduces the risks substantially and bring down costs.
- Keep IT “Available”
The OES is fully fault tolerant with no single point of failure. Each hardware layer is architected with redundancy which offers extreme high availability to the customers.
- Consolidate IT Environments
With OES, you can consolidate the various workloads spread across in multiple places in your data center to a single box and save in administrative, real estate costs.
- Simplify Cloud
Oracle Claim that, the OESs are simplified form a cloud which offers availability, flexibility, scalability and performance.
“Why OES?” – The Reality
Cost Reduction and Lower Risks
This is a “tricky” one-liner from Oracle and one should be extremely careful when you take and digest it upfront. At a high level, Oracle claim that, since the engineered systems are highly powerful, eliminate many overheads such as node-to-node traffic by consolidating them to a high performance inter-connection within a single rack AND/OR interconnected racks, the actual compute and storage needs are “much lesser” compared to conventional infrastructure individually bought and built together. But, a decision maker should keenly work on the following before convinced on this point
- What is your actual compute and storage need NOW and approximate requirements for the same year on year for coming 3-5 years?
- What is the estimated cost for these requirements when you deploy the systems in conventional hardware? The estimated cost MUST include, individually procured infrastructure , space / cooling/power requirements, associated license requirement (including Oracle) and all associated support costs
- What is the translated requirements to Oracle Engineered System requirements Year on Year? Keep in mind that, OES comes as “Building Blocks”; i.e it can start with a block like a 1/8th Rack and can be augmented only in similar blocks (if you have that in on-premise. In cloud, the story is different; so is the price which we will discuss later)
- What is the cost of OES year on year for the requirements plotted in C? DO NOT miss out to ensure “SUPPORT PART” including the much highlighted PLATINUM support by Oracle Sales. (The default platinum is nothing short of a dismal break-fix kind of support).
- What is the performance benchmarks on conventional hardware Vs OES? Document it at granular level. Do it realistically and thoroughly. This is really important and if you miss this out this part, you may regret it later and wonder what went wrong.
- Sum up the requirements for conventional hardware year on year and that of OES year on year. If the delta requirements for your requirements are abysmal or relatively low, then, forget about adopting Oracle Engineered systems. Not only that you will not get the ROI, you will be struggling with many other management overheads including capacity upgrades, idle hardware for prolonged time period etc.
- Cost Comparison and Decision
- The support cost for OES : Initially the Oracle Sales person would pitch you with “Platinum” service which is “in-built” in OES. But be aware that, that Platinum service is nothing but to enable Oracle to break-fix the integral/critical parts of the system “when it crashes”. They claim that, they will know before or when it crashes and fix it. But mostly, it will be, you, the customer who will know it at first. Then, once you are fairly comfortable with taking the system, there will be an extended offer over and above Platinum which comes for a price. Add that cost to your total cost to manage OES.
- Sum up the costs for both conventional and OES separately and yearly. Ensure to consider all the associated costs including facility costs and rentals, license costs and required resource costs to manage it. Please bare in mind that, though Oracle claim that, it requires lesser resources to manage an Engineered System, the bitter fact is that, it requires extremely high expertise to manage it in-house (even if you don’t touch the internal things in the Box). Those resources, even if small in number , are costly. If you give it out to Oracle, that’s a top-up cost which will be a considerable number.
- The inference : You would be able to arrive at the decision by comparing the data collected through this method. It is purely depending up on your “actual requirements” which is closely linked to the business plan. I am not saying that, it will never beneficial to anyone. But in general, unless you have a mammoth workload which exponentially grow year on year, the true ROI and reduction in TCO will take its sweat time ; for a much larger pain.
- Probably, with this much data in hand, this will be a juncture you may think of to migrate your workload to public cloud. Please watch out this series for more on this.
Sethunath U N
This is Part 1 of Many in this series which discuss about Oracle Engineered Systems. If you think of considering Oracle Engineered Systems adoption or want to evaluate your decision, do contact us at email@example.com
Image Courtesy : Oracle Corporation