ENKI versus Amazon EC2

What's the difference between ENKI and Amazon EC2?

While we are accustomed to competing with Amazon EC2, we also understand that the two systems (ours and theirs) are quite different and comparison isn’t straightforward, since they were designed for different purposes.  Amazon aims to provide virtual hardware, while we provide virtual information technology solutions. Ultimately, we are focused on providing overall value to our customers that ensures their business success rather than exactly matching a particular resource rate which may not offer them overall value in the long term - though we're often more than competitive. Our offering is designed for customers who want the infrastructure transparency and reliability that VMWare offers together with no-compromises hardware systems, while receiving that service in the context of an available fully-managed enterprise class IT operations solution, even if they may not use full management at first. Our PrimaCloud solutions are clearly not for everyone just as Amazon isn't, and we’re not attempting to be the low price lead, but rather provide a guaranteed-to-work solution for business critical applications at leading total cost of operations (TCO). Perhaps most importantly of all, in ENKI you have a personal connection with a partner interested in your success, and flexible enough to find ways to help you when you need help, or provide an excellent alternative to building your own IT capability.  ENKI's strength is not just technology, but relationship.

We understand that comparing ENKI prices to Amazon is difficult due to differences in packaging, pricing, measured quantities, and hidden charges.  Amazon's per-instance rates may be lower that ENKI's per-GB based pricing in some cases, but due to Amazon's complex charges for bandwidth, storage, I/O operations, IP addresses and other resources as well as its highly variable performance that forces customers to overbuy compute resources, many of our customers have saved more than 20% over deploying with Amazon.  The best way for you to compare is to let us create a custom quote for you that takes into account your actual compute needs and architecture.

What's the difference between an ENKI "core" and an Amazon EC2 "core"?

ENKI''s cores are the equivalent of a 2.8GHz Opteron or Xeon; Amazon has listed theirs as a 1GHz Opteron or Xeon, 2006 vintage. This means that one ENKI core is 3-4 Amazon ECUs. ENKI cores come with 4GB of RAM, which can be allocated across all your instances are you desire.

Can you compare ENKI's PrimaCloud and Amazon EC2 with respect to technology?

ENKI was designed from the ground up to eliminate the type of performance problems that have been documented with Amazon EC2 and other cut-rate cloud computing solutions, which can result in up to a tenfold variability in performance over the course of the day.   In particular, storage speed and interconnect speed are critical to high-performance cloud computing.  ENKI has addressed storage speed with high-speed SAN systems that are monitored to prevent oversubscription and have optional SSD acceleration for database-intensive loads.   Our interconnect is Infiniband, which offers 40 times the performance of Amazon's 2Gb/s connections, allowing servers and storage to work together without bottlenecks during peak loads.   Think of Amazons technology as a slice of a dedicated server, with slow (GbE) access to shared resources.  ENKI's approach is to provide fully virtualized high-performance storage, compute and interconnect, and let you decide how much of each you need for your instances.

Can you compare ENKI's PrimaCloud and Amazon EC2 in detail?

 

	Amazon AWS	ENKI PrimaCloud
Reliability	SLA compensation for 99.95% uptime requires user request and only amounts to 10% credit per month. Amazon EC2 and S3 have experienced several severe failures just this year, and more last year, effectively giving their customers a 10% discount in those months, if they apply for it.	10x money-back guarantee for unplanned outages exceeding 99.99% uptime per month.  Guarantee is paid automatically, based on our automatic monitoring (if deployed), and is applied at month-end.  We also measure uptime as whether your operating system is up, not just the virtual machine.  The intention is to assure you that we have skin in your game.  "We're up when you're down."
Security	Amazon instances are open to a variety of attacks.  Advanced protection requires installation of firewall and antivirus software on each instance, which soaks up significant CPU time, which is charged for.	All instances have multiple layers of protection including a hardware firewall.  Active Intrusion Detection/ Intrusion Prevention firewalls with antivirus available to protect against hacking, viruses and DDOS attacks.  ENKI's Virtual Private Data Center paradigm means that your instances communicate with each other on secure private networks that are not visible to other customers or from the Internet, setting you free to choose whatever security is appropriate, applied at the gateway to your VPDC.
Performance	According to independent studies, Amazon performance can vary up to 10x over the course of the day, due to customers competing for shared resources.  ENKI's incoming customers have reported interconnect delays between Amazon instances (and to EBS) of over 100 milliseconds.  This delay produces I/O wait that wastes your computing power, effectively costing you up to 10 times more than the listed price per unit of compute power.	PrimaCloud attacks the two main sources of performance problems with high-performance architecture.  I/O congestion, in which virtual machines compete for access to interconnect, is addressed with dual Infiniband interconnects to each server for a total of 80Gbit/second bandwidth to other servers and storage.   Storage oversubscription, in which mulitiple tasks compete for storage throughput, is managed by using dedicated SANs for each cluster, with optional SSD acceleration to reduce access times to under a millisecond.  Monitoring measures storage access times to ensure they don't slow your application down.
Pricing	Pricing starts at $0.10 per hour for the smallest instance (with Linux only). Additional "hidden" and difficult-to-estimate charges for storage accesses, both input and output bandwidth, IP addresses, and other consumables produce bills that can be surprisingly large and defy estimation based on published rates. Savings on reserved resources are contingent on buying them for a year in advance. Amazon does offer free or micro-instances, but our customers report that they are not suitable for production computing.	ENKI offers customer-defined instances sizes, priced $0.086/GB-Hr of RAM at 8GB of usage, which includes approximately 6 Amazon ECUs.  Steep discounts off of base price up to and beyond 60% for increasing usage levels over a term period or for prepayment. Increasing discount levels are applied automatically as your usage goes up so you know you're getting the best discounts.  You don't have to pay in advance - ever!   Storage and bandwidth prices also drop automatically with increasing usage. Automatic workload management can save you up to 80% over purchasing static resources by scaling your application automatically to match the load from users/customers - without having to manage your application or use complicated APIs. Significant savings are available on reserved resources (paid for monthly instead of in advance) or on prepaid instances, with savings of up to 75% possible.
Instance configuration	Fixed ratio of RAM to CPU with limited instance configuration choices requires paying for resources you don't use, much like with physical servers.  Because of variable performance, you have to overbuy on instance size.	User-defined ratio of RAM to CPU on a per-instance basis.  Charges are made only on total resource allocations allowing balancing different CPU/ RAM ratios across multiple instances resulting in overall cost savings.  You can change this ratio on each instance as your needs change.
Persistence	No guaranteed instance storage persistence: you must provide for application failover and redundancy since instance storage is lost on failure or shutdown.  Essentially instance storage is only for scratchpad or temporary use since it can disappear at any time.  External Elastic Block Store slows access to critical persistent data. Extra instances are required to provide failover capability, increasing cost.  You will need to redesign your application to work around these limitations.	Instances are guaranteed to be persistent.  PrimaCloud and The Computing Utility closely model a highly reliable physical data center using VMWare technology or Xen technology.  VMWare or AppLogic handle server and application failover automatically, restarting failed instances on new hardware without data loss - guaranteeing persistence.  This is accomplished by storing all data off of the server so that none is lost on failure.
Granularity/Scalability	You must purchase resources in increments of one of Amazon's limited set of instance sizes, which may result in increased costs. New instances are available 5-15 minutes after request - if resources are available.  Our customers have reported that Amazon "runs out" at inconvenient times.	ENKI instances can be adjusted to any resource allocation from 1% to 100% of the physical server. ENKI instances are guaranteed to be available either if customer reserves them with a small monthly fee equal to 1/6th list price, or if ENKI is notified at least two weeks in advance of peak loads.
Compute Power	Amazon measures resources in terms of approximately 1 GHz 2006-era "cores."	ENKI measures resources in terms of cores equivalent to a 2.8GHz processor, nearly 4 times the power.
IP addresses	Dynamic - you have to code around changing IP addresses.  IP address remaps are charged.  Fixed IP addresses at $7.20/mo	Static.  5 IP addresses are included with each VPDC.  IP addresses beyond 5 cost $1/mo.
Controllability	Instances and storage are allocated by API call or web interface. Your application is responsible for scaling itself unless you can use a generic scaling algorithm.   Third party management tools cost extra.	ENKI adjusts your VPDC for you based on your requests or self service via our online portal which makes configuring virtual datacenters easy. We offer root logins to virtual servers in your VPDC, as well as management plans that eliminate the need to log in or know anything about how to manage the server. Your VPDC can scale automatically in response to changing load conditions - according to your rules.
Observability	You provide your own monitoring if you want to go beyond just reporting on server status to measure application status.	ENKI offers and supports enterprise-grade monitoring at the application and server levels, as well as remote user-experience monitoring
Locality/Predictability	You don't know where your application is running: there is little control over deployed data center architecture.  Large scale storage only available at near-WAN speeds.  Latency between instances and to S3 can vary from 10s to 100s of milliseconds, with large jitter.  Local resource zones only reduce latency and jitter, don't eliminate it.	ENKI manages its own data centers and controls locality of data and compute functions.  Large scale storage is available at enterprise-style direct-connect speeds, and all instances of your VPDC are on a local network with microsecond latencies, perfect for transaction oriented or telecommunications applications.  (Not applicable to geographically distributed applications.)
Bandwidth Pricing	You are charged for both input and output bandwidth.	ENKI charges only the max of input and output bandwidth, resulting in significant savings for storage-intensive applications.  Prices start at $0.17/GB and drop rapidly with increased usage.
Bandwidth Quality	No specification.  Highly variable due to random distribution of Amazon's data centers	ENKI's data centers are located at carrier meet-me points such as MAE West, with carefully selected carrier mix to ensure minimum ping and transit times.  You may find that with one running instance at ENKI, you are getting equivalent latency to many instances running in different Amazon data centers.
Deployment	You figure out how to deploy, architect, and manage your application.  You have to understand Amazon's architecture to deploy and manage your application.  Many users hire additional staff to do this, negating most of their cloud savings.	Do it yourself through our GUI or standard VMWare API, or let ENKI's experienced staff of IT professionals deploy a virtual private data center of your design, with your software in it, using our available full management services.
Vendor Lock-In	Amazon. AWS uses a proprietary virtual machine image structure. If you want to move your application, you have to reinstall it.	Running applications including an entire VPDC can be moved to another provider's datacenter since ENKI's services are based on industry standard of VMWare.  In addition, our portal can schedule automatic conversion of other hypervisor images (V2V conversion).
Support	Forums and user groups.  Paid support plans available, but only for Amazon's technology, not your applications and deployment, so you still need Amazon administrators on-staff.	You call the phone and we answer, with guaranteed service response times.   Optionally, our system admin experts take responsibility for keeping your application up and running at a fixed per-instance-month charge.  We will support almost any software running on your machines, even your own software (with suitable training.)  ENKI's high-context support, which we call "Operations Services" acts like your own IT department, with a dedicated representative that can execute longer-term projects on your infrastructure for you, or understand ongoing problems.
Relationship	Amazon's business model is a hands-off vending machine for virtual hardware.  Relationship is not important to Amazon, so you don't have an IT partner, just a compute vendor.	ENKI offers Virtual IT services, necessitating a collaborative partnership with its customers that allows it to serve as its customers' IT department.  Relationship is a core value at ENKI and we get 50% of our business through referrals as a result.
Intended Audience	Do-it-yourself IT professionals, short-term resource consumers, deployments without performance requirements, organizations that have more free personnel time than cash.	Organizations requiring enterprise-grade guaranteed performance and service levels without staffing investment, wishing to take advantage of ENKI's extensive data center design expertise including creating the infrastructure for NetSuite.  Startups/entrepreneurs, Enterprise business units/intrapreneurs, ISVs, lean/agile organizations, companies requiring enterprise virtual IT services.