Performance Expert Services
Our scientific approach to performance analysis and engineering is proven. We work with top 5 Investment Banks, Telcos, and technology vendors. If you have a hardware, software, or network performance problem that is impacting your business, contact us to arrange a meeting.
Performance is money. IT performance matters, whether you are running reports, processing transactions, or simply serving content to impatient users. Performance impacts the effectiveness and productivity of your business. When performance problems happen, chasing them down is one of the most difficult, most draining activities there is. As soon as you try to measure performance on a running system, you introduce overhead and make the performance problem worse.
Moore's Law predicts a doubling of semiconductor performance every 2 years. Your hardware may be refreshed every 4. When was the last time you noticed a 400% improvement in performance? The reality is that underlying performance improvements are smothered by inefficiency, dependencies, bottlenecks, and increased complexity. More often than not hardware refreshes simply hide performance problems until the next upgrade is due.
There is no one single technology or product to save you. Not multi-core, cloud computing, solid-state storage, or NoSQL. Some of these technologies if applied without sufficient understanding can even make performance problems worse. As vendors add intelligence to their silicon and firmware, Operating Systems get larger and more complex. Layers of code and abstraction multiply, scope grows for obscure interactions that result in performance problems.
Our work begins when performance problems defeat your team’s efforts, and are beyond the scope of vendor patches and support contracts. We look across systems, technology layers, and vendors. We unlock latent performance and remove processing bottlenecks.
Our specialty is achieving a transformation in performance when there is no budget to reach it by conventional means.
Although the individual technical details of every performance engagement differ, most follow a similar path. Our experts begin by gaining a thorough understanding of the business impact of the problem you are experiencing. We ascertain what your IT staff have attempted already, and what action has been taken by vendors and suppliers. We take a scientific approach to performance.
The Scientific Method
- Measure & hypothesise, build a model for ideal performance
- Test & verify findings, compare the real system to the ideal model
- Take corrective action & repeat, until observed performance approaches the ideal
We carry out research, simulation, and live testing to produce a vendor-neutral report. We detail the precise nature of the problem and recommend actions to alleviate it. Our consultants are then able to put those actions into practice on-site or hand over to your staff for implementation. For projects where implementing the recommendations may take several months, we are able to monitor, guide, and intervene directly if required. Our engineers are engaged to solve a wide variety of IT performance challenges, including:
- In-place Tuning & Optimisation
- High Performance Infrastructure Engineering & Design
- Infrastructure Lifetime Extension
- Vendor Benchmarking & Evaluation
- Performance Crisis Management & Mitigation
Improving the performance of your Information Technology isn't just something to be done when the business complains. Better performance often means more productive staff, greater business flexibility, and happier customers. In some cases a leap forwards in IT performance can mean strategic competative advantage, even dominance of a market. Each industry has it's own story of how faster processing has precipitated a minor revolution. Financial markets moved from batch to real-time views of risk, retail from passive "shop windows" to per-customer "suggestions" , Telecoms from fixed plans to dynamic "friends and family" billing. Performance is profit.
Whatever your aspirations, improved system performance can provide very practical benefits;
- Delay capital spending on new equipment
- Next years hardware performance now
- Free up staff time spent nursing overloaded systems
- Resolve multi-vendor, multi-layer blame loops
- Maximise value from sunk investments
- Reduce license, energy, and hosting costs via higher utilization, and fewer systems
Consider This Service When
- You need increased performance but have no capex budget
- A critical application is hindered by poor performance
- A legacy system is a bottleneck to your processing
- You require confirmation of a vendor's advice on an upgrade
- IT performance prevents business flexibility
- You can't meet your backup and maintenance windows
50% of engagements begin after a hardware upgrade fails to deliver expected performance.
70% of engagements find a solution without further capital equipment investment.
5 months is the average amount of time clients waste on their problem before calling us.
10 is the average number of days taken to analyse, diagnose, and present mitigation steps.
5 minutes is all it takes to get in touch with us.
High-Definition Video Content Distribution
Out client was a provider of on-line training solutions for the automotive sector. They were struggling to deliver high quality streaming video to a network of several hundred dealerships throughout Europe from their UK datacenter. After receiving pricing from Content Delivery Network providers, and an alternate quote for more datacenter bandwidth, they contacted 360is to see if there was a more cost effective way of achieving their goal.
During a 2 week engagement 360is consultants designed a private Peer-To-Peer network, using in-place datacenter and dealership assets (servers and desktops with DSL connections) for the distribution of high quality videos. No additional hardware was required. No more bandwidth was needed. There was no licensed software bought. No increase in operating expenditure at either the datacenter or the dealerships. This solution provided the distribution capability the client required, saved them several hundred thousand pounds per year of CDN and Internet access costs which would otherwise have been incurred.
40 CPUs Work From An 8 CPU System
Our client was a large international bank. They had built a sophisticated system for distributing patches, packages, user-credentials, and configuration files securely across their global estate. They had 7000+ UNIX hosts and a strict rolling maintenance window.
The system had become overloaded, it struggled to complete its task within the time window. It was frequently still sending files at start-of-day. The hardware had been upgraded to 8 CPUs and could not be upgraded further. Functions could not be easily moved to a larger platform or divided up due to complex interdependencies with other systems. Migration to an entirely new system would take months and would be extremely risky. The in-house solution was superior to alternative commercial products on the market, and free of per-host license costs. 360is were retained for 2 weeks to analyse and solve the problem.
Consultants determined that the process was network and CPU constrained. Our changes were made to the system and CPU consumption dropped by 80% and network traffic reduced. This allowed for increasing demand, bought the client several years of spare capacity and required no capital or operational expenditure. There was no disturbance to the complex software running on the system nor any changes to the rest of the UNIX estate.
Virtualizing Difficult Workloads
Our client was one of the worlds best known multinational brands. They were were moving their IT estate to a virtualised platform. Although the migration had worked well for most systems, a particular workload was performing very poorly when virtualised. The problem was characterised by long CPU spikes at 100% and unacceptable interactive response for users of the system during these spikes. After 6 months work by their own IT team, and that of their strategic IT partner, they were no closer to finding a solution. 360is was contacted by one of the vendors involved.
Over a period of several days 360is consultants analysed the application, Operating System, and underlying virtualisation platform. We determined empirically where performance was being lost. The deliverable was a report, containing hard data to be presented to the vendors responsible for the problem components. It detailed options for improving performance by 20% while the cause (a kernel-space routine in the hypervisor) was addressed. The report also contained details of the more radical step of swapping out one software component for another, which would provide a 100% improvement, with no adverse impact on the business critical application user.
Storage Area Network Under Performance
Our client was a fast growing online company. They had invested heavily in their first enterprise grade SAN. After only 6 months the system was suffering from declining performance and the software developers were talking about going back to local RAID storage. The SAN reseller immediately advocated the purchase of more equipment. Our client was unwilling to invest further capital without rigorous justification. Independent advice from 360is, backed up by detailed performance data couched in business terms, allowed our client to justify further investment in their infrastructure. The original SAN capacity had simply been exhausted through rapid growth in the business.
Our scientific approach to performance analysis and engineering has been proven in previous engagements. We work with top 5 Investment Banks, Telcos, and technology vendors. If you have an IT performance problem that is impacting your business, contact us to arrange a no-obligation meeting with one of our consultants.